Isoform-Specific Effects of Wild-Type Ras Genes on Carcinogen-Induced Lung Tumorigenesis in Mice

The gene KRAS is commonly mutated in lung cancer to encode a constitutively active and oncogenic protein that is well established to initiate and maintain lung tumorigenesis. However, the remaining wild-type KRAS protein, or the other family members HRAS and NRAS, can still be activated in the presence of oncogenic KRAS. Moreover, loss of any one of these three genes has been shown to increase the sensitivity of mice to the carcinogen urethane, which induces Kras mutation-positive early lung lesions. To determine the contribution of progressively disrupting Hras and Nras genes on urethane lung tumorigenesis, mice with different combinations of wild-type and null alleles of Hras and Nras were exposed with urethane and tumor burden was assessed. As previously reported, loss of one allele of Hras increased the sensitivity of mice to this carcinogen, and this effect was further exacerbated by the loss of the second Hras allele. However, loss of one or both alleles of Nras failed to alter tumor burden, either in the absence or presence of Hras, after exposure to urethane. Additionally, no obvious difference between lung lesions in mice with wild-type versus null alleles was detected, suggesting that wild-type Ras proteins may exert a tumor suppressive effects at the time of initiation, although other interpretations are certainly possible. In summary, these data suggest that in some genetic backgrounds inactivation of different wild-type Ras genes can have different effects on urethane-induced lung tumorigenesis.


Introduction
Approximately one third of human cancers have an activating mutation in one of the three RAS genes; HRAS, NRAS, or KRAS [1]. RAS proteins function as molecular switches, alternating between inactive GDP-bound and active-GTP bound forms [2]. Conversion to the active GTP-bound state is greatly accelerated by guanine nucleotide exchange factors (GEFs) that promote the exchange of GDP for GTP, whereas GTPase activating proteins (GAPs) enhance GTP hydrolysis, reverting RAS proteins back to their inactive GDP-bound state [3].
Oncogenic mutations in these genes, typically at codons 12, 13, or 61, inactivate the inherent or GAP-stimulated GTPase activity of the enzyme, thereby rendering RAS constitutively active. In this state, chronic activation of downstream signaling pathways, including MAPK, PI3K, and RalGEF, results in uncontrolled cellular proliferation among other effects, driving tumor formation and growth [4].
KRAS is the most commonly mutated isoform in human cancers, and is well known to both initiate and maintain a wide spectrum of cancers, but especially pancreatic, colorectal, and lung cancers [5]. In regards to the latter, approximately 220,000 individuals in the United States are diagnosed with lung carcinoma each year, making it the second-most commonly diagnosed type of cancer [6]. Non-small-cell lung carcinoma (NSCLC) is the most prevalent type of lung cancer, and KRAS is the driving oncogene in about 20 to 30% of these cases [7]. As such, NSCLC is the cancer in which RAS mutations affect the highest number of individuals. Therefore, it is important to understand the role that RAS signaling plays in this disease.
Although oncogenic RAS proteins are well appreciated to drive tumorigenesis, the remaining wild-type family members can also be activated in the presence of the oncogenic proteins [8][9][10][11][12]. The consequences of this activation are complex. In some cases, particularly in models of early stage cancer, the wild-type isoforms inhibit oncogenic RAS-driven tumorigenesis [8,[13][14][15]. In contrast, in KRAS-mutation-positive cancer cell lines, wild-type RAS signaling instead promotes proliferation and tumorigenesis [9][10][11][12][13]16]. These observations suggest that wild-type RAS proteins may have temporal, tissue, or other context-dependent effects on oncogenic RAS-driven tumorigenesis.
With regards to lung cancer, loss of both alleles of wild-type Hras, loss of both alleles of wild-type Nras, or loss of the remaining wild-type Kras allele, significantly increases the number of oncogenic Kras-driven lung tumors in mice exposed to the carcinogen urethane [13,15,17]. In agreement, the opposite experiment, namely increasing the expression of the remaining wild-type Kras allele, reduces lung tumorigenesis driven by oncogenic Kras [18]. These findings suggest that in early lung tumorigenesis induced by the carcinogen urethane, the wildtype Ras genes have tumor suppressive properties.
Since the loss of any of the wild-type Ras genes individually enhances urethane carcinogenesis, we explored whether progressive loss of an increasing number of Ras alleles would have an additive effect at promoting lung tumorigenesis. To this end, we generated littermates with different combinations of Nras and/or Hras null alleles. These mice were then treated with the carcinogen urethane to induce Kras mutation-positive lung lesions. We report here that, consistent with previous findings [13], loss of one allele of wild-type Hras enhanced lung tumorigenesis in mice treated with the carcinogen urethane, and that this effect was more pronounced when both alleles of Hras were inactivated. Interestingly however, loss of one or both alleles of wild-type Nras had no significant effects on the number of lung tumors that developed. Moreover, loss of wild-type Nras did not enhance the numbers of tumors observed in an Hras-null background. These results suggest the intriguing possibility that there are isoform-specific effects of wild-type Ras genes on urethane-induced lung tumorigenesis in some genetic backgrounds.

Generation of mice and genotyping
Hras -/and Nras -/mice were obtained from the NCI and Eugenio Santos (University of Salamanca, Spain). Hras +/-;Nras +/mice on a mixed 129/BL6 background were bred to obtain littermates with various combinations of native and null Hras and Nras alleles. Genomic DNA isolated from a total of 268 pups was genotyped for Hras and Nras wild-type and null alleles as previously described [19].
Urethane carcinogenesis 9 to 14 mice per cohort at 7 to 9 weeks of age were intraperitoneally injected with 1 mg/g urethane (Sigma-Aldrich, dissolved in sterile PBS) as previously described [15]. Mice were then visually monitored for signs of pain, distress, or moribundity (sudden behavioral change, poor posture or ambulation difficulty, loss of hair coat condition, sudden activity change, painful facial expressions, neurological disorders, and cardiopulmonary disorders, according to the Duke University Institutional Animal Care and Use Guidelines) and weighed three times a week. Animals reaching these endpoints, or exhibiting at 15% reduction in weight were to be euthanized. However, no mouse exhibited signs of pain, distress, or moribundity or during the course of evaluation, and as such all mice in this study (barring one animal that was found dead for unexplained reasons) were humanely euthanized at the fixed time point of 11 months post-injection. Humane euthanasia was achieved by gradually replacing atmosphere air with CO 2 followed by bilateral thoracotomy. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and under protocol A279-13-11 approved by the Institutional Animal Committee on Use and Care at Duke University.

Quantification of lesions
Mice were euthanized at a fixed endpoint of 11 months after administration of urethane, the lungs were removed, and the number and size of surface tumors visible under a dissecting microscope at 10X magnification was documented without knowledge of the genotype. The left lung from each mouse was formalin-fixed, paraffin-embedded, and sectioned every 200 μm. Eight such sections were mounted on slides and H&E stained for pathologic analysis. Two pathologists (JMC and ENP), blinded to genotype, independently graded and counted the lesions in each section from 9 to 14 mice per cohort.

Immunohistochemistry
Additional slides prepared from the above sectioning of the left lungs of mice exposed to urethane were subjected to epitope retrieval and stained with, anti-Ki67 (Thermo Scientific RM9106), anti-CC3 (Cell Signaling, D175), anti-P(Thr 202/Tyr 204)-Erk1/2 (Cell Signaling 4376), and anti-P(Ser473)-Akt (Cell Signaling 4060), anti-F4/80 (ABDserotec MCA497GA), and anti-CD3 (ThermoFisher RM9107) antibodies, followed by peroxidase-based detection with Vectastain Elite ABC Kits (Vector Labs) and counterstained with haematoxylin. Photographs were taken of 6 to 10 (10X) random fields in a blinded fashion. The tumor areas were determined from these photographs using the freehand selection tool in Image J (imagej.nih.gov). Color thresholding was applied to determine positive-staining areas, using the same parameters for each tumor image. Areas staining positive by these parameters were selected and the positive-staining area in pixels recorded. The percentage of positive-staining area was calculated by dividing the positive-staining area of the tumor in pixels by the total area of tumors in pixels.

Statistics
Statistical Analyses were performed using GraphPad Prism v5 (GraphPad Software). A 2-sided, unpaired t-test was used to compare the number of lesions and levels of immunohistochemical staining between cohorts, and p values were calculated using the log-rank (Mantel-Cox) test.
As expected, urethane induced lung lesions in all the mice (e.g. Fig 1B). Consistent with previous studies [13], microscopic analysis revealed that Hras +/-;Nras +/+ mice, which have one Hras null allele, developed roughly three times more visible tumors than control Hras +/+ ;Nras +/+ mice. This effect was further exasperated by loss of the second wild-type allele, as Hras -/-;Nras +/+ mice developed, on average, more than twice as many tumors as Hras +/-;Nras +/+ mice (Fig 1C). In contrast to the progressive effect that the loss of wild-type Hras alleles had on lung tumorigenesis, there was no statistically significant difference in the number of lung tumors in Hras +/+ ; Nras +/or Hras +/+ ;Nras -/- (Fig 1D) mice that contain either one or two Nras null alleles. Thus, in this mixed genetic background, the progressive loss of Hras alleles, but not Nras alleles, results in an increase in tumor burden upon exposure to the carcinogen urethane.

Concomitant loss of wild-type Hras and Nras does not enhance the formation of urethane-induced tumors
Previous studies have shown that loss of both wild-type Hras and wild-type Nras individually enhanced urethane-induced tumorigenesis in a dosage-dependent manner [13]. This suggests that loss of both of these isoforms could have an additive effect on promoting tumorigenesis in this model. To explore the effects of losing both Hras and Nras on urethane-induced tumors, the number of lung lesions that developed in double homozygous Hras -/-;Nras -/knockout mice was determined. These mice developed approximately the same number of tumors as Hras -/-;Nras +/+ mice lacking only Hras (Fig 2A). Thus, the enhanced tumorigenesis in mice lacking wild-type Hras is unaffected by the additional loss of wild-type Nras in this genetic background.
To assess if loss of progressively more Ras alleles alters tumor progression, tumors developing in Hras +/+ ;Nras +/+ control mice, Hras -/-;Nras +/+ mice lacking Hras, Hras +/+ ;Nras -/mice lacking Nras, or Hras -/-;Nras -/mice lacking both Hras and Nras were graded in a blinded fashion by 2 pathologists. All but one of the lesions were either atypical adenomatous hyperplasia (AAH) or adenoma (AD), both of which are early stages of lung tumorigenesis. One mouse (Hras -/-;Nras +/+ ) had a lesion graded as adenocarcinoma (not shown in figure). Comparing tumors from all 4 genotypes revealed no significant difference in the grade of lesions between these different genotypes (Fig 2B). Thus, loss of any of the tested combinations of wild-type Hras or Nras alleles did not alter the grade of lesions developing in the urethane model of lung tumorigenesis.

Loss of wild-type Hras and Nras does not affect levels of proliferation, apoptosis, or downstream-signaling within urethane-induced tumors
To explore the effect of loss of wild-type Ras genes on tumor characteristics, 6 to 10 lung lesions from urethane-treated mice with the two extreme genotypes, namely Hras +/+ ;Nras +/+ versus Hras -/-;Nras -/-, were stained for markers of apoptosis (cleaved caspase 3), proliferation (Ki67), inflammation (F480), and Ras signaling (phosphorylated ERK and AKT). No significant differences in immunohistochemical staining using antibodies against cleaved caspase 3 (Fig 3A), Ki67 (Fig 3B), or F480 ( Fig 3C) were observed between these two cohorts. Furthermore, the average levels of phosphorylated AKT ( Fig 3D) and ERK (Fig 3E) immunostaining within the lesions were also similar between these two cohorts. Similarly, loss of just Hras had no measurable effect of MAPK signaling, as assessed by the levels of phosphorylated ERK (S1A Fig). Loss of wild-type Hras and Nras in combination increases the number of lymphoid aggregates in the lungs of urethane-treated mice Pathological examination of lung sections revealed an aggregation of lymphoid cells in mice exposed to urethane (Fig 4A). To assess whether disruption of either or both Hras and Nras alleles influenced this lymphoid aggregate phenotype, two pathologists determined the number of lymphoid aggregates in each section of lungs isolated from the 4 genotypes Hras +/+ ; Nras +/+ , Hras -/-;Nras +/+ , Hras +/+ ;Nras -/-, and Hras -/-;Nras -/-. Binning these data into groups of no aggregates, 1-5 aggregates, or greater than 5 aggregates per section revealed that the number of lymphoid cell clusters were higher in Hras -/-;Nras +/+ mice, and even more prevalent in Hras -/-;Nras -/mice (Fig 4B). We note here this was not associated with changes in F480 staining (Fig 3C). Staining for cluster of differentiation (CD3) was also similar between Hras +/+ ; Nras +/+ , Hras -/-;Nras +/+ , and Hras -/-;Nras -/genotypes (S1B Fig). Thus, loss of wild-type Hras enhances the lymphoid response in the lungs to the chemical carcinogen urethane, and this response is further augmented by the additional loss of wild-type Nras.

Discussion
We demonstrate that loss of wild-type Hras enhances carcinogen-induced lung tumorigenesis in mice, in agreement with previous findings [8,13]. This effect was not, however, observed in mice lacking wild-type Nras, nor did loss of Nras increase the sensitivity of Hras -/mice to urethane. Such a finding presumably reflects a difference in either the expression or function of these two Ras isoforms. We note that loss of Nras had previously been shown to yield more lung tumors in mice treated with urethane [13]. One possible explanation for this difference is that the experiments described here employed a mixed genetic background while the aforementioned study use a pure 129 strain. Indeed, strain background is well established to greatly influence the susceptibility of mice to urethane-induced tumorigenesis [20]. More specifically, the mix background of the study described here was less sensitive to urethane carcinogenesis, which may mask effects due to disrupting Nras alleles. Nonetheless, in this background loss of Hras promoted Kras-driven lung tumorigenesis, while loss of Nras had no measureable effect. Levels of Ras signaling, as measured by pAKT and pERK immunostaining, were similar in the lesions between mice with wild-type versus null Hras and Nras alleles. These lesions also had similar staining patterns for markers of proliferation, apoptosis, and inflammation. It is therefore possible that the effect of losing wild-type Hras occurs before the tumors develop, perhaps at the stage of initiation, and as such would not manifest a phenotype in established lesions. For example, loss of wild-type Hras may alter lung development, increasing the number of the cell types susceptible to urethane carcinogenesis. Alternatively, losing Hras in the stroma could affect tumorigenesis without altering tumor cell signaling. Finally, we recognize that there may yet be differences in signaling or markers of cellular features like proliferation, apoptosis or inflammation, but the effects are just too subtle to detect by immunohistochemical staining. Further studies are needed to determine the exact mechanisms by which loss of Hras promotes the growth of carcinogen-induced lung lesions.
Finally, lymphocyte infiltration was observed in Hras knockout mice exposed to urethane, a phenotype more pronounced upon the additional loss of the Nras gene. Aggregates of lymphoid tissue in the lung are most often in response to infection, but occasionally are seen in cases of NSCLC [21]. Knockdown of wild-type Nras has also been shown to alter pathways related to immune responses [22], and Hras -/-;Nras -/mice are unable to mount an immune response to parasitic infection, due to hampered Th1 immunity [23]. Taken together, these results suggest a relationship between the loss of wild-type Ras and an immune response, although whether this is related to urethane carcinogenesis, and if so how, remains to be determined.
In summary, we demonstrate that different wild-type Ras isoforms can have different effects on urethane carcinogenesis, adding important insight into the relationship between wild-type Ras proteins and early tumorigenesis.