An Autochthonous Model of Lung Cancer Identifies Requirements for Cellular Transformation in the Naked Mole-Rat

Abstract

There is growing interest in understanding the mechanisms underlying differences in cancer incidence among species (comparative oncology). The naked mole-rat (NMR) is often referenced as “cancer-resistant” and prior studies focused on identifying mechanisms explaining this. However, efforts to assess this in vivo have been limited. Herein, we provide evidence that the NMR presents as a novel autochthonous model of lung tumor initiation, driven by an introduction of the oncogenic Eml4-Alk fusion protein using CRISPR-mediated genome editing. Whereas in mice the inversion alone is sufficient to drive tumorigenesis, the inversion alone was insufficient to drive tumorigenesis in the NMR lung and tumor development required additional losses of the tumor suppressors p53 and pRb. Our findings suggest that the proposed “resistance” of the NMR to the development of cancer may reflect that the genetic events leading to tumor initiation are likely to be comparable to those present in human cells.

Introduction

Naked mole-rats (NMRs, Heterocephalus glaber) have emerged as a popular model for the study of cancer resistance and aging(1,2). This spike in popularity can be attributed, in part, to their long-lived lifespan and seldom reported cases of malignant disease(3-7). Over the past decade, NMRs have further cemented their reputation as a cancer-resistant species through a plethora of studies proposing multiple biological mechanisms that could contribute to their longevity and healthy lifespan(8). These mechanisms are varied, including early cell contact inhibition(9), enhanced DNA repair mechanisms and genome stability(10-12), high tolerance to oxidative stress(13,14), and reduced inflammatory responses(15). Despite the steady increase in NMR-focused studies, many of these proposed mechanisms are limited to in vitro analyses and have yet to be truly tested in a physiological context through rigorous in vivo studies. Additionally, to date, there are no genetically engineered models of disease established in NMRs. To address these shortcomings, we sought to develop a first of its kind endogenous cancer model in the NMR.

Evidence of spontaneous tumors in NMRs has been previously reported(5-8). Therefore, we hypothesized that NMRs would develop tumors, given the appropriate initiating events. While the specific biological processes that may drive tumorigenesis in the NMR are yet to be determined, we hypothesized that NMRs may have different requirements for cellular transformation compared to other species. For example, it is well established that there are central differences in cellular transformation of rodent cells compared to human cells. In mice, the overexpression of one or two cooperating oncogenes is enough to lead to aberrant cell growth and subsequent transformation. In contrast, oncogene activation alone is not sufficient in human cells. Additional events are required, including the loss of cell cycle and cell death regulators, such as the tumor suppressors p53 and pRb, and the expression of telomerase(16-19).

To test our hypothesis, we developed a model of lung cancer using CRISPR-Cas9 to induce an oncogenic inversion, which would result in a Eml4 and Alk fusion protein. This chromosomal inversion is present in 4–6% of human lung adenocarcinoma (LuAD) cases and is a strong oncogenic driver in mice(20-22). In a previously established Eml4-Alk mouse model, the induction of the translocation by CRISPR-Cas9 was sufficient to drive significant tumor development with 100% penetrance within three months of infection(21). We utilized this approach to generate the Eml4-Alk inversion in NMRs and assess tumorigenesis in the lung.

Results

Method validation in naked mole-rats

Given that an in vivo NMR model has not been developed to-date, the assessment of CRISPR-Cas9 functionality and suitability of an adenoviral-based delivery approach is critical. The Eml4 and Alk alleles are located on the same chromosome in both mice and NMRs (Supplementary Fig. S1a). We therefore adapted the dual-guide CRISPR-Cas9 strategy used in the murine model for editing of the NMR genome (Supplementary Fig. S1b).

To first assess the suitability of an adenoviral-based delivery system, we dual-infected NMR lungs (n=3) with two adenoviruses containing GFP (Ad-GFP) or RFP (Ad-RFP) via intra-nasal instillation. We collected the lungs 72 hours post infection and analyzed the dissociated cells via flow cytometry, to ensure the dual-infection leads to a sufficient subset of double positive NMR lung cells (Supplementary Fig. S2a). We also confirmed that mice are susceptible to double infection with adenovirus when infected with Ad-GFP and Ad-RFP (n=3; Supplementary Fig. S2b). In tandem, we tested the CRISPR-Cas9 adenovirus in NMR cells. We infected immortalized NMR skin fibroblasts(23) (ISF) in vitro or NMR lungs in vivo (n=2) with the Eml4-Alk CRISPR virus (Ad-EA) and collected the cell or lungs 72 hours post infection. Genomic DNA was extracted and recombination assessed by PCR. This analysis confirmed that these infections led to the successful generation of the Eml4-Alk inversion in ISFs and NMR lungs (Supplementary Fig. S2c-d).

The Eml4-Alk inversion generates a functional fusion protein

After confirming that the designed genomic inversion event occurred, we next confirmed whether the presence of the Eml4-Alk fusion gene led to any phenotypic changes associated with the presence of this oncogenic driver, such as increased cell proliferation. Towards this goal we generated a set of NMR skin fibroblast cell lines, both primary (PSF) and immortalized (ISF), that express the Eml4-Alk mRNA fusion, as confirmed by RT-PCR (Supplementary Fig. S3a). The growth of the Eml4-Alk cell lines (Eml4-Alk⁺) was then compared to control cells. The presence of the Eml4-Alk inversion resulted in a significant growth advantage for the Eml4-Alk⁺ compared to control cells, as shown in both PSF and ISF (Supplementary Fig. S3b-c). Treatment with the FDA-approved Alk inhibitor crizotinib impaired this growth advantage in the Eml4-Alk⁺ immortalized cell lines, suggesting that the Eml4-Alk fusion contributes to a growth advantage in these cells (Supplementary Fig. S3d).

Eml4-Alk does not induce tumorigenesis

Knowing that the Eml4-Alk fusion protein provides a growth advantage in vitro, we next sought to determine if this effect persists in vivo. NMRs were next infected intra-nasally with Ad-EA (n=7) or a control CRISPR virus (Ad-con; n=4). Lungs from both infection groups were collected at three months post-infection and preserved via FFPE processing. The three-month timepoint for analysis was chosen based on previous results in the mouse model in which all animals infected with an inversion-inducing adenoviral vector displayed lung adenocarcinoma(21). We confirmed this result in-tandem with our NMR infections, using mouse Ad-EA (Ad-mEA) on C57BL6 mice and found similar results (n=10; Supplementary Fig. S4a-d).

Upon examination, control NMRs did not exhibited any evidence of tumor development as visualized by H&E (Supplementary Fig. S5a-b). H&E staining showed normal histology consistent with healthy lungs. The Eml4-Alk infections in the NMR recapitulated what we observed in the control infections, displaying no evidence of pre-cancerous pathology at a three-month timepoint (Supplementary Fig. S5c-d). To determine if more time is necessary for tumorigenesis, an additional group of NMR Ad-EA infections (n=5) were carried out for 6 months, again with no evidence of tumor development (Supplementary Fig. S5e-f).

We extracted DNA from FFPE sections from the three-month infections and analyzed them via genomic PCR (Supplementary Fig. S6). From these samples, none of the control infections had detectable levels of the Eml4-Alk inversion, whereas a subset of NMR lungs from the Ad-EA infections did present with detectable levels of Eml4-Alk (3 out of 7 animals tested). These data suggest that the presence of the Eml4-Alk inversion alone is not sufficient to induce tumorigenesis in NMR lungs.

Loss of Tp53/Rb1 does not induce tumorigenesis

Unlike in mice, where the expression of a strong oncogenic driver such as activation of Kras or Eml4-Alk fusion protein can lead to extensive tumor burden in the lung, the induction of the Eml4-Alk inversion did not lead to tumorigenesis in NMRs. To further investigate how additional mutations affect NMR tumor initiation, we decided to assess two additional driving events: the loss of the tumor suppressors p53 and pRb. Tp53 and Rb1 mutations are common co-alterations in ALK⁺ NSCLC patients(24,25). Additionally, Tp53 and Rb1 mutations are sufficient to drive tumorigenesis in small cell lung cancer (SCLC) models in mice(26-28).

To test the effects of knocking out both Tp53 and Rb1, we developed a CRISPR-Cas9 adenovirus to target both simultaneously (Ad-PR). This adenovirus was modeled after the same dual-guide CRISPR system we used for Eml4-Alk, with one guide RNA (sgRNA) targeting Tp53 and one for Rb1, to create inactivating mutations and/or indels in coding exons. The ability of the vector to generate these mutations was confirmed by infecting NMR PSF with Ad-PR, collecting DNA after 72 hours, PCR amplifying the target region, and sequencing (Supplementary Fig. S7a-b). Following this confirmation, NMRs were infected with Ad-PR through intranasal instillation and assessed 6 months post-infection, a timepoint chosen based on the latency of tumor development in the lungs of p53/Rb knockout mice(26-28). Lungs infected with Ad-PR showed no signs of tumor development at a six-month timepoint (Supplementary Fig. S8a-b).

In addition to the Ad-PR vector, we made a set of dual guide adenoviruses individually targeting only Tp53 (Ad-p53-d) or only Rb1 (Ad-Rb1-d), which result in large genomic deletions of the respective genes (Supplementary Fig. S9a-b). The purpose of this was to make identification of Tp53 and Rb1 mutations easier by following the same PCR-based identification system we have used for the Eml4-Alk inversion. We assessed the virus in vivo at 72-hour post-infection to confirm that these deletions are detectable in NMR lungs (Supplementary Fig. S10a). The Ad-p53-d or Ad-Rb1-d CRISPR viruses were then used in dual-infections with the Ad-EA virus to determine if combinations of Eml4-Alk/Tp53 or Eml4-Alk/Rb1 mutations would lead to tumor formation. The validation that NMR lungs would have a sufficient percentage of cells that are double-positive upon infection with two adenoviruses was previously confirmed (Supplementary Fig. S2b). As was the case with the Eml4-Alk only infections, the NMR lungs appear normal and tumor-free at a 3-month timepoint (Supplementary Fig. S10b-e).

The combination of the Eml4-Alk inversion and loss of Tp53/Rb1 leads to tumorigenesis

Upon confirming that combinations of Eml4-Alk expression and mutations in either Tp53 or Rb1, or the combination of Tp53 and Rb1 did not lead to tumorigenesis, we assessed whether the combined effect of all three events could drive tumor development. To address this, we performed intra-nasal instillation on NMRs (n=10) with the Ad-EA and Ad-PR viruses (referred to as Ad-PREA infection). From these infections, a subset of our NMRs (3 of 10) developed significant tumor growth at around a 15-week timepoint. The morphology of the tumors was consistent with pleomorphic carcinoma, mostly comprised of adenocarcinoma with areas of spindle cell and giant cell carcinoma (Fig. 1a-e).

Figure 1. Histology and IHC of NMR lungs from dual-infection group with significant tumor burden.

NMR lungs from infections were collected, preserved through FFPE, and processed for H&E and IHC staining. Representative images of lungs from the NMR dual infection group (Ad-EA/Ad-PR). (a-e) H&E staining. (a) Overview of NMR lung. Star = airway; circle = normal lung tissue; arrows = adenocarcinoma nodules. 25X magnification. (b-c) Solid adenocarcinoma nodules. Tumor cells are large and round with pleomorphic nuclei, prominent nucleoli, and large cytoplasms. (b) 200X magnification. (c) 400X magnification. (d) Pleomorphic spindle/fusiform cells in a carcinomatous embolism. 400X magnification. (e) Giant tumor cells with polymorphic and multi-nucleated nuclei with immune cell infiltration. 400X magnification. (f) SPC. 200X magnification. (g) CK14. 200X magnification. (h) NKX2-1. 200X magnification. (i) Ki-67. 200X magnification.

IHC staining was then carried out using antibodies that were validated for their ability to detect NMR proteins (Supplementary Fig. S11a-d). The IHC displayed largely heterogenous staining, with distinct areas of positive staining for SPC (alveolar type II (AT2) marker), CK14 (epithelial cell marker), and NKX2-1 (diagnostic marker for LuAD) (Fig. 1f-h). Most tumors stained strongly with Ki-67, indicating the tumors were highly proliferative, with an estimated Ki-67 index of 80% (Fig. 1i). These data further support the pathological classification of the NMR lung tumors as pleomorphic carcinoma. As is seen in human pleomorphic carcinoma, our tumors are likely of epithelial origin but have areas of dedifferentiation into an epithelial-mesenchymal transition.

Mutational analysis of these tumors was conducted by isolating individual tumor regions using laser capture microdissection (LCM) of FFPE tissues (Fig. 2a). DNA was extracted from these tumors and used for genomic PCR to detect the Eml4-Alk inversion or targeted deep sequencing of Tp53 and Rb1 to detect mutations or small deletions (indels). This analysis revealed that all the tumors from which DNA was successfully isolated contain the Eml4-Alk inversion (Fig. 2b). Furthermore, deep sequencing analysis confirmed that all isolated tumors contained mutations in Tp53 and Rb1 (Fig. 2c). The most common Tp53 and Rb1 mutations, which were identified in 100% of the tumors, were disruptive frameshift mutations which have a high probability of leading to an inactive protein. Most tumors contained additional mutations in these two genes which could further disrupt function. The mutational profiles of these tumors, along with the results of the previous infection experiments, suggests that all three genetic events are necessary to drive formation of tumors in NMR lungs.

Figure 2. LCM strategy and subsequent mutation analysis.

(a) Whole-slide images of the NMR lungs with significant tumor burden. Tumors were isolated as shown by the green lines. Number indicates tumor identification. Images were scanned at 20X magnification. (b) DNA was extracted from LCM-captured tumors. Genomic PCR analysis shows the presence of the Eml4-Alk inversion. Tumor identification numbers are indicated above the gels. WT indicates a matched wildtype tissue control (infected with Ad-PREA, no tumor). Low and high exposures shown. (c) Top 20 mutations identified in Tp53 and Rb1 with an allelic frequency of 0.01 or greater. The percent frequency of each mutation is indicated on the left Y-axis. Individual tumors are shown on the X-axis, color coordinated by the NMR the tumor was taken from: NMR-PREA1, blue; NMR-PREA2, purple; NMR-PREA3, red. Red boxes indicate which mutations are present in the indicated tumors. Mutation type is indicated on the far right.

Single-cell RNA-sequencing of NMR tumors

To characterize the cellular milieu of the NMR lung tumors, we performed single-cell RNA-sequencing (scRNA-seq) of lung tumors from three NMRs. For each NMR, all lung tumors were removed and pooled into one sample for analysis. The total number of tumors from each NMR ranged from 1–7 tumors per lung. NMRs were infected with Ad-PREA (3x10¹⁰ pfu). UMAP projections depict the various cellular populations in the NMR tumors (Fig. 3a). NMR tumors show high proportions of macrophages and plasma cells, with low proportions of B cells (Fig. 3a and Supplementary Fig. S12). We also observed a low percentage of alveolar type I (AT1) and AT2 cells (Fig. 3 and Supplementary Fig. S12), likely due to limited recovery of these cells with the extraction protocols favoring non-adherent cells. AT1 and AT2 cells are of particular interest, given the identity of the cell of origin for pleomorphic lung carcinoma is currently unknown(29).

Figure 3. Single-cell RNA-sequencing of NMR lung tumors.

(a) Combined global UMAP projection of NMR lung tumor cells (n=3). Clustering of cells from the individual NMR are shown to the right. (b) Clustering of the T cell population based on the following markers: CD2, CCL5, CD3, CD3G, CTSW, CXCR3, LAT, LCK, NKG7, RAC2, SKAP1, tnfrsf18. (c-g) Expression of specific T cell markers. (c) CD3E = T cell surface glycoprotein CD3 epsilon chain precursor. (d) CD3D = T cell surface glycoprotein CD3 delta chain precursor. (e) CD4 = T cell surface glycoprotein CD4 (f) CD8 = T cell transmembrane glycoprotein CD8 (g) NKG7 = natural killer cell granule protein 7. (h) GZMA = granzyme A. (i) CCL5 = Chemokine (C-C motif) ligand 5.

In addition to lung epithelial cells, we were interested in the adaptive immunity of the NMR lung tumors. Recent studies have shown that NMRs have a unique repertoire of T cells with a large proportion of γ T cells and a lack of canonical NK cells. The NMR tumors had a high proportion of T cells and were the second largest cell population. In concordance with previous studies, we could not identify canonical natural killer (NK) cell populations in NMRs (Fig. 3b-g). The T cells in the cluster expressed NK markers (Nkg7, Gzma), but this overlapped with other markers including Cd3e, Cd3d, and Ccl5 (Fig. 3h-i).

Spatial analysis of NMR lung tumors

We sought to confirm specific cell types in the NMR tumors and confirm several markers from the scRNA-seq data. We utilized BaseScope Duplex technology to assess localization of cell-type specific markers and the Eml4-Alk inversion using NMR-specific probes. In all NMR lungs, the expression of Eml4-Alk was restricted within the lung tumors (n = 10; Fig. 4a-d and Supplementary Fig. S13a-h). Eml4-Alk expression co-localized to a remarkable extent with the AT2 epithelial cell marker Sftpc, (Fig. 4e-g and Supplementary Fig. S14a-f), suggesting that these tumors may have an AT2 origin(26).

Figure 4. Spatial analysis of the Eml4-Alk fusion expression in NMR tumors.

Images are from NMR #42394 and representative (n=10). NMR lungs were collected, fixed, and paraffin embedded. Unstained slides were cut at 5 um and used in BaseScope Assay (see methods for details). Slides were scanned with an Aperio slide scanner at 40X. Images were taken on a Zeiss Upright FL microscope. (a) Overview image of scanned slide. Red indicates regions with Eml4-Alk expression (tumor regions). Boxes indicate regions shown in b-d. (b-d) 400X images of tumor border regions. Dashed lines indicate approximate tumor borders. Red dots indicate Eml4-Alk. (e-f) Red dots indicate Eml4-Alk and green dots indicate Sftpc positive cells in lungs from 2 NMRs. The images were segmented by cell using QuPath software and quantified manually for either Eml4-Alk, Sftpc, or both, represented in Venn diagrams (g).

Markers for T cells (Cd3e), B cells (Cd79a), and macrophages (C1qa) were present in the tumors as well, with a higher percentage seen in tumor tissue than in adjacent normal tissue (Supplementary Fig. S15a-f). Expression for these immune cell markers generally did not overlap with Eml4-Alk expression.

Discussion

As previous studies suggested that NMRs are resistant to the development of tumors, we implemented an in vivo model using an approach previously used to generate a model of Eml4-Alk driven lung adenocarcinoma in mice. We utilized CRISPR-mediated genome editing to induce this inversion in somatic cells of the NMR lung. Our data indicate that while we can generate the designed inversion, which results in expression of an active Eml4-Alk inversion, this alone was insufficient to drive tumorigenesis in the NMR lungs. This is in contradiction to what is seen in mice, where Eml4-Alk alone leads to rapid, extensive tumor growth(21). Based on these findings, we hypothesized that NMRs likely require additional “hits” to drive cellular transformation in vivo, as suggested to be the case in human cells.

Human cells require more than the activation of an oncogene to lead to cellular transformation. Additional events include the loss of cell cycle regulators, cell death regulators, and the expression of telomerase(16,18). To assess if NMRs had similar transformation mechanisms, we targeted the tumor suppressor genes Tp53 and Rb1. We first assessed these alone or in combination and found that, similar to the Eml4-Alk infections, infections with our Ad-PR to induce Tp53 and Rb1 mutations also did not lead to tumor development. Furthermore, the combinations of Eml4-Alk/p53 or Eml4-Alk/Rb1 mutations did not lead to tumorigenesis. Previous literature has suggested that NMR cells can be transformed in culture with the simultaneous loss of p53 and pRb(9). This has been contradicted in a subsequent study that observed cellular crisis and a cessation of growth with loss of p53 and pRb(30), which is consistent with our in vivo data. We have demonstrated here that the loss of these critical tumor suppressors fails to lead to tumor development. Similarly, using the combination of the expression of an oncogene and loss of a single tumor suppressor we were unable to observe any lesions in the NMR lungs, suggesting the combinations of these events are insufficient in inducing tumorigenesis. Tumorigenesis was only observed in NMRs infected with the combination of the Eml4-Alk inversion and both Tp53 and Rb1 mutations. Overall, our model had a penetrance of around 30%. This could be due to the inability of the dual infections to target the relevant tumor initiating cells simultaneously, enhanced responses to DNA damage through DNA repair, and/or immune pathways. Further studies are needed to address further proposed NMR cancer resistance mechanisms.

The tumors that developed in the NMR lungs were large and heterogenous and likely indicative of late-stage disease. Importantly, the tumors were identified as pleomorphic lung carcinoma, a rare subtype of NSCLC for which no mouse models are currently available. Given the rarity of this form of NSCLC, there is a paucity of studies with published data with a limited number of patients, and obtaining access to these has proved challenging. However, there is evidence that the majority of pleomorphic lung tumors have a mutation in Tp53, along with additional driver mutations such as Kras, EGFR, and Alk fusions (all identified as Eml4-Alk fusions), with Alk mutations being relatively rare(31,32). The driving mutations are generally mutually exclusive, an observation consistent with other subtypes of NSCLC(33). Mutations in Rb1 were also uncommon, but not mutually exclusive with the Tp53 alterations(31,32). Future comparative analysis using human data sets, when available, would be beneficial in identifying similarities between the NMR tumors and human disease.

We identified and interesting distribution of cell types in NMR lung tumors, including high proportions of macrophages, plasma cells, and T cells. We saw relatively small populations of cells identified as AT1/2 cells when employing scRNA-seq analysis. This low percentage likely reflects a bias towards non-adherent cells in our methods when preparing the single cell suspension. The suboptimal number of cells made further downstream analysis of potential Alk, Tp53, and Rb1 signaling signatures difficult. We are currently working on developing improved methods to capture these AT cells specifically, so that these mutational analyses can proceed. We will use this data to determine whether AT cells serve as the cell of origin of our model. Knowing the cell of origin can reveal additional similarities in tumor initiating events between NMRs and humans and provide further information regarding tumor heterogeneity and potential responses to therapies.

In regards to immune cell components in the lung TME, we find a high abundance of macrophages and plasma cells in NMR tumors, suggesting roles for both the innate and adaptive immune system in this model. Investigating the immune microenvironment of NMR tumors is of particular interest and will be the subject of future studies, for a couple of reasons. Recent analyses of human pleomorphic lung carcinoma has shown high rates of infiltrating CD8+ T cells and in some cases, tumor-associated macrophages (TAMs)(29,34). Additionally, recent analyses of the NMR immune system have identified interesting features, such as a lack of canonical NK cells and the high abundance of γδ T cells(35). Our exploratory preliminary analysis of the immune system using scRNA-seq allowed us to confirm some of these features, such as the lack of NK cells, and set up analysis pipelines for further studies of NMR tumor immunology.

Through spatial analysis and species-specific probes, we were able to further investigate the genetic background of our tumors and validate some of the scRNA-seq findings. We identified tumor-specific expression of Eml4-Alk, which co-localized with the AT2-specific marker Sftpc. This further confirms that our lung tumors are likely of epithelial origin, like many other NSCLC tumors. We also found increased expression of immune cell markers for T cells, B cells, and macrophages in the tumors, which were mutually exclusive from cells expressing Eml4-Alk.

Herein, we have not only provided evidence of critical requirements for tumorigenesis in NMRs, but also generated a genetic model for pleomorphic lung carcinoma that fills a much-needed gap. Importantly, we have shown that NMRs may more closely recapitulate human tumor initiation and may provide a better representation for tumor initiation. Continuing studies on NMRs and the generation of additional in vivo models with different driving events will be essential for confirming this hypothesis. Our observations do not necessarily discredit the claims that NMRs are resistant to cancer, but rather confirm a difference in the susceptibility to development of cancer compared to other rodents such as mice and provide evidence that NMRs undergo tumorigenesis through mechanisms akin to those observed in humans. An important next step is distinguishing whether this susceptibility is impacted by variations in cell autonomous versus non-cell autonomous functions. We observed expression of important immune components, including T cells, B cells, plasma cells, and macrophages, both intratumorally and tumor adjacent. Whether these components specifically play a role in the unique tumor initiating events observed in the NMR remains to be determined and will require the further development of species-specific reagents.

The naked mole-rat is certainly an interesting model system and demonstrates promise in mirroring human cancer development. However, it takes significant resources to establish an NMR breeding colony. NMRs are not as prolific of breeders as mice, so achieving high numbers of animals for in vivo studies takes more time than mice. Despite these challenges, NMRs are worth studying, due to their potential as a highly accurate model system for human cancer and their unique biological systems.

Methods

Cloning.

Guide RNAs (sgRNAs) were cloned into a variety of CRISPR plasmids, including pSpCas9(BB)-2A-Puro (PX459) V2.0 (a gift from Feng Zhang, Addgene plasmid # 62988 ; http://n2t.net/addgene:62988 ; RRID:Addgene_62988)(36), pX333 (Addgene plasmid # 64073 ; http://n2t.net/addgene:64073 ; RRID:Addgene_64073)(21), and Adeno Cas9 (Addgene plasmid # 64072 ; http://n2t.net/addgene:64072 ; RRID:Addgene_64072)(21). sgRNAs sequences are listed in Supplementary Table 1 in the supplemental methods. Plasmids were digested in CutSmart Buffer (NEB) with either BbsI (NEB) or BsaI (NEB). sgRNAs were annealed using T4 Polynucleotide Kinase (PNK, NEB). Annealed sgRNAs were ligated into the digested plasmid using Ligation Mighty Mix (Takara). Ligation products were transformed into TOP10 competent E. coli (Invitrogen). Positive clones were maxi-prepped with the ZymoPure II Plasmid Maxiprep Kit (Zymo Research). All reactions were carried out according to the manufacturer’s instructions.

Adenovirus infections.

In vitro –

NMR cells were plated 500,000 cells in a 10cm dish 24 hours prior to infection. Unless otherwise stated, 10μl of adenovirus was used for infections (1 x 10⁶ pfu – 1x10⁸ pfu depending on adenovirus titer). Infection media was left for 72 hours before being replaced by fresh media. Cells were then either collected for analysis or left to proliferate.

In vivo –

Infection mix was prepared using MEM, 1mM CaCl₂, and 3 x 10¹⁰ pfu total of the selected adenovirus(es) to a total volume of 50μl. Infected NMRs were non-breeders between ages 1 to 3. NMRs or C57BL6 mice (RRID:IMSR_JAX:000664) were placed under anesthesia (isoflurane) and infected via intranasal instillation in two doses, 25μl each. Following infections, NMRs were observed until recovery and placed back in their appropriate colonies. Infections with a control virus (Ad-con) confirmed there were no animals that displayed an adverse response or developed detectable pathological manifestations as a consequence of adenovirus exposure. In all cases, equal numbers of males and females were used.

Genomic PCR.

Genomic DNA was extracted using the DNA Mini Kit (Qiagen). FFPE DNA was extracted using the GeneRead FFPE DNA Kit (Qiagen). All PCRs were done using AmpliTaq Gold 360 PCR Master Mix (Invitrogen). PCRs were run as 50μl reactions according to manufacturer’s protocol. PCR products were run on 1% agarose gels at 100V for 30 – 40 mins or sent for Sanger sequencing (Eton Bioscience and Genewiz). PCR primers are listed in Supplementary Table 1 in the supplemental methods.

RT-PCR.

RNA was extracted using the RNeasy Mini Kit (Qiagen). cDNA was transcribed from total RNA using the iScript cDNA Synthesis Kit (Bio-Rad). 1μg RNA was used for each reaction, unless starting material was scarce, where the maximum amount of RNA was run based on available material. PCRs were run as previously described, using 1μl of cDNA from the RT. RT-PCR primers are listed in Supplementary Table 1 in the supplemental methods.

Growth assay.

Cells were seeded in 24-well culture plates at 10,000 cells per well (unless otherwise specified), in triplicate, for a 5-day assay (15 wells total). Cells were manually counted with a hemocytometer at the same time every day. Media was changed every other day. Significance was calculated using mixed effects analysis or two- or three-way ANOVA, depending on the number of experimental and treatment groups (indicated in figure legends). Multiple comparison analysis was also conducted to compare individual experimental conditions. Calculations were conducted using Prism version 7.

Histology.

Lungs were harvested at the specified time points, inflated by intratracheal injection with 10% formalin (for fixation) or PBS (for immediate analysis), incubated for 24h in 10% formalin, and then transferred to 70% ethanol for another 24h. Lungs were then embedded in paraffin for future analysis. For staining, tissue slices were cut from FFPE blocks and mounted on microscope slides. Lungs from all NMRs were stained with hematoxylin and eosin (H&E). Lungs were also processed for a variety of markers with immunohistochemistry (IHC). Antibodies and dilutions used for this project are listed as follows: rabbit anti-Ki67 (CST, 1:100), mouse anti- CK14 (Abcam, 1:400), rabbit anti-SPC (Millipore-Sigma, 1:2000), and mouse anti-NKX2-1 (Novus Biologicals, 1:1000). Antibody details are listed in the Supplementary Methods (Supplementary Table 2). All slides were scanned at 20X magnification using the Leica Aperio AT2. 2.3.3. We observed that NMR lungs may be more fragile than mouse lungs when it comes to inflation. Our H&E stains often show ruptured airways and alveolar tissue, along with ruptured blood vessels that lead to blood cells ubiquitously observed amongst the alveolar sacs. This is merely due to the collection process and not a consequence of the adenovirus infections.

Laser capture microscopy.

Laser capture microscopy (LCM) was completed using the Arcturus XT Laser Capture Microdissection System. Individual tumors were identified from the scanned H&E slides. An unstained, 10μm thick section of FFPE tissue was mounted on PEN membrane slides (2μm; Leica). Once tissue was cut, forceps were used to pick out dissected tissue from the slide and place in 1.5mL microcentrifuge tubes for DNA or RNA extraction.

Targeted deep sequencing.

Library preparations –

Following PCR amplification (primers listed in Supplementary Table 1 in the supplemental methods), amplicons were sequenced using the Illumina Nextera XT kit following the manufacturer’s protocol (Illumina). Briefly, following the quality control screening of 100ng of each amplicon on a 1% agarose gel, both amplicons for each sample were pooled equimolar and 1ng of input DNA was used for the Nextera tagmentation reaction, followed by PCR amplification and sample indexing. The libraries were purified, quantitated with the Qubit Fluorometer (ThermoFisher Scientific), and screened on the Agilent TapeStation 4200 using the TapeStation High-Sensitivity D1000 Screen Tape (Agilent Technologies). The final libraries were normalized, denatured, and sequenced on the Illumina MiSeq sequencer using a Nano flow cell to generate approximately 20,000 150-base read pairs per sample. Data analysis methods provided in the Supplementary Methods.

Spatial Analysis with BaseScope Assays.

Preparation –

Slides for BaseScope assays were prepared using FFPE lung tissues. 5μm slices were cut and placed unstained on plus-charged slides. Paraffin removal, antigen retrieval, and protease digestions were conducted according to manufacturer’s protocol. Briefly, slides were baked at 60°C for 1 hour. Paraffin was removed with two xylene washes (5 mins, room temperature), two 100% ethanol washes (2 mins, room temperature), and dried at 60°C for 5 mins. Hydrogen peroxide (provided) was added and incubated for 10 min at room temperature. Slides were rinsed in pre-warmed water before being placed in pre-warmed 1X RNAscope target retrieval buffer in a steamer and steamed for 15 min. Slides were rinsed in water then incubated in 100% ethanol for 3 min at room temperature and dried for 5 min at 60°C. RNAscope protease III was added to the slides and incubated for 30 min at 40°C. Slides were washed in water.

Probe hybridization and amplification –

Probes were added to prepared slides. For single-plex, probe was added directly (~4 drops). For duplex, probes were added in a ration of 1:50 of probe C2:C1. Hybridization occurred at 40°C for 2 hours. Slides were washed twice in 1X RNAscope wash buffer (provided) and left in 5X SSC buffer overnight before amplification. Amplification steps were carried out according to manufacturer’s protocol, with the optional extension steps added. All probes are listed in Supplementary Table 3 in the Supplementary Methods.

Single-Cell RNA-Sequencing.

Sample preparations –

Lungs were inflated with 1X PBS and collected from NMR. Tumors were removed from normal tissue manually using forceps and scissors. Tumors were dissociated into single cells as described in the supplementary methods. Cell suspensions were washed twice with 1X PBS containing 0.04% BSA.

Library preparations and sequencing –

Single-cell RNA sequencing was conducted using the 10X Genomics Chromium X/IX (10X Genomics, Pleasanton, CA) at the Molecular Genomics Core of the Moffitt Cancer Center and Research Institute. Following wash steps, cells were resuspended in the same buffer (1X PBS with 0.04% BSA) according to the 10X Genomics cell preparation protocol. Cell viability and counts were assessed using AO/PI dual fluorescent staining and analyzed with the Nexcelom Cellometer K2 (Nexcelom Bioscience LLC, Lawrence, MA). Subsequently, cells were loaded onto the 10X Genomics Chromium Single Cell Controller at a concentration of 1,000 cells/μl, aiming to encapsulate 10,000 cells per sample. Individual cells, reagents, and 10X Genomics gel beads were encapsulated into nanoliter-sized Gel Beads in Emulsion (GEMs). Reverse transcription of poly-adenylated mRNA occurred within each droplet at 53°C. cDNA libraries were prepared in a single bulk reaction following the 10X Genomics Chromium NextGEM Single Cell 3’ Reagent Kit v3.1 guidelines, and 50,000 sequencing reads per cell were generated using the Illumina NovaSeq 6000 instrument. Data processing, including demultiplexing, barcode processing, alignment, and gene counting, was carried out with the 10X Genomics CellRanger v8.0.0 software, and results were visualized using the 10X Genomics Loupe Browser v8.0. Data analysis methods are provided in the Supplementary Methods.

Data Availability Statement

Raw data from the single-cell RNA-sequencing experiment and lung tumor deep sequencing were deposited in GEO repository (accession number GSE305565 and GSE305802, respectively).

Animal studies

All and animal studies and care requirements were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC).

Statement of significance.

To identify evolutionary divergent mechanisms of cancer resistance we assessed tumorigenesis in vivo using the naked mole-rat, a species considered to exhibit cancer resistance. Our findings suggest that the proposed “resistance” of NMRs to the development of cancer may reflect tumor initiation mechanisms comparable to the mechanisms present in humans.