We describe a method to significantly enhance orthotopic engraftment of lung cancer cells into the murine lungs by pre-conditioning the airways with injury. This approach may also be applied to study stromal interactions within the lung microenvironment, metastatic dissemination, lung cancer co-morbidities, and to more efficiently generate patient derived xenografts.
Lung cancer is a deadly treatment refractory disease that is biologically heterogeneous. To understand and effectively treat the full clinical spectrum of thoracic malignancies, additional animal models that can recapitulate diverse human lung cancer subtypes and stages are needed. Allograft or xenograft models are versatile and enable the quantification of tumorigenic capacity in vivo, using malignant cells of either murine or human origin. However, previously described methods of lung cancer cell engraftment have been performed in non-physiological sites, such as the flank of mice, due to the inefficiency of orthotopic transplantation of cells into the lungs. In this study, we describe a method to enhance orthotopic lung cancer cell engraftment by pre-conditioning the airways of mice with the fibrosis inducing agent bleomycin. As a proof-of-concept experiment, we applied this approach to engraft tumor cells of the lung adenocarcinoma subtype, obtained from either mouse or human sources, into various strains of mice. We demonstrate that injuring the airways with bleomycin prior to tumor cell injection increases the engraftment of tumor cells from 0-17% to 71-100%. Significantly, this method enhanced lung tumor incidence and subsequent outgrowth using different models and mouse strains. In addition, engrafted lung cancer cells disseminate from the lungs into relevant distant organs. Thus, we provide a protocol that can be used to establish and maintain new orthotopic models of lung cancer with limiting amounts of cells or biospecimen and to quantitatively assess the tumorigenic capacity of lung cancer cells in physiologically relevant settings.
Lung cancer is the leading cause of cancer related deaths worldwide1. Patients with lung cancer eventually succumb from metastasis to distant organs, notably to the central nervous system, liver, adrenal glands, and bones2,3,4. Thoracic malignancies have been traditionally classified as small cell lung cancer (SCLC) or non-small cell lung cancer (NSCLC)5. NSCLC is the most frequently diagnosed malignancy and can be subdivided into different histological subtypes, including lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC)6. Genomic analysis of resected human primary lung cancers has revealed that tumors within a given histotype can also express diverse molecular perturbations, further contributing to their divergent clinical progression and confounding patient prognosis. The remarkable heterogeneity of lung cancers represents a significant challenge to the rational design, pre-clinical testing, and implementation of effective therapeutic strategies. Consequently, there is a need to expand the repertoire of tractable experimental lung cancer models to study the diverse cellular origins, molecular subtypes, and stages of this disease.
Various approaches using animal models have been employed to study lung cancer in vivo, each with their own advantages and disadvantages depending on the biological question(s) of interest. Genetically engineered mouse models (GEMMs) can target specific genetic alterations in a given progenitor cell type, resulting in tumors that progress within an immunocompetent host7. While extremely powerful and clinically relevant, the latency, variability, and/or lung tumor morbidity associated with GEMMs can be prohibitive to certain quantitative measurements and the detection of late stage metastasis in distant organs8. A complementary approach is the use of allograft models, whereby lung cancer cells, obtained either directly from a mouse tumor or derived first as established cell lines in culture, are re-introduced into syngeneic hosts. Analogously, lung cancer xenografts are established from human cell lines or patient derived tumor samples. Human cell line xenografts or patient derived xenografts (PDXs) are generally maintained in immunocompromised mice and therefore preclude complete immune-surveillance9. Despite this drawback, they provide an avenue to propagate limiting amounts of human biospecimens and study fundamental in vivo properties of human cancer cells, which encode for more complex genomic aberrations than GEMM tumors.
One useful property of allografts and xenografts is that they are amenable to traditional limiting cell dilution assays, employed to quantify the frequency of tumor initiating cells (TICs) within a malignant cell population10. In these experiments, a defined number of cells are injected subcutaneously into the flank of animals and the frequency of TICs can be estimated based on tumor take rate. Subcutaneous tumors however can be more hypoxic11 and may not model key physiological constraints of the lung tumor microenvironment. Intratracheal delivery of epithelial stem or progenitor cells into the lungs of mice is a method to study pulmonary regeneration and airway stem cell biology12. However, the engraftment rate from this technique can be relatively low, unless the lungs are first subjected to physiological forms of injury, such as viral infection13,14. Support from inflammatory stromal cells and/or the disruption of the lung basement membrane may improve retention of transplanted cells into relevant stem cell niches in the distal airways15. Fibrosis inducing agents can also pre-condition the lungs to enhance engraftment of induced pluripotent cells16 and mesenchymal stem cells17. Whether similar forms of airway injury can affect the engraftment rate, tumor initiating capacity, and outgrowth of lung cancer cells has yet to be systematically assessed.
In this study, we describe a method to increase the efficiency of orthotopic lung cancer cell engraftment, by pre-conditioning the lungs of mice with injury. LUAD arises in the distal airways with a significant subset of these cancers developing a fibrotic stroma18 that often correlates with poor prognosis19. Bleomycin, a natural nonribosomal hybrid peptide-polyketide, has been extensively utilized to induce pulmonary fibrosis in mice20. Airway instillation of bleomycin first promotes epithelial attrition in the alveoli and recruitment of inflammatory cells, including macrophages, neutrophils and monocytes21. This is followed by tissue remodeling in the distal airways, basement membrane reorganization22,23 and extracellular matrix (ECM) deposition24. The effects of a single bleomycin injection are transient, with fibrosis resolving after 30 days in most studies25. Using both allograft and xenograft models, we tested if pre-conditioning the airways of mice with bleomycin could significantly increase the take rate of LUAD cells in the lungs.
All experiments were carried out in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) at Yale University.
1. Set Up / Preparation of the Reagents.
2. Bleomycin Treatment
3. Monitoring Mice Post-Intubation
4. Engraftment of Lung Adenocarcinoma Cell Lines.
NOTE: Perform engraftment of cells 14 days after the injection of bleomycin (step 2.1).
5. Monitoring of Tumor Growth by Bioluminescence Imaging
6. Tissue Isolation and Processing.
To increase the efficiency of LUAD cancer cell engraftment into the lungs of mice, we developed a protocol that first pre-conditions the airways using bleomycin followed by orthotopic tumor cell injection (Figure 1). We confirmed that even when administered into immunocompromised athymic mice, bleomycin induced transient fibrosis by day 14 as evidenced by loss of airway architecture and increased collagen deposition (Figure 2). Gross fibrosis in these mice resolved 50 days post bleomycin injection (Figure 2; right panels) consistent with prior studies24.
Based on these observations, we engrafted LUAD cells of different origins into the lungs of various mouse strains that had been pre-conditioned with a single dose of vehicle or bleomycin 14 days prior. For example, a suspension of LUAD cells from the established human LUAD cell line, H2030, was delivered intratracheally into the lungs of athymic mice. After 35 days, mice pretreated with bleomycin had high lung tumor burden as detected by bioluminescence (Figure 3A). Conversely, in vehicle-treated animals, no tumors were detected over the same time frame (Figure 3A). Lung tumor burden was confirmed by histology following necropsy. Lungs pretreated with vehicle had no evidence of tumor nodules whereas bleomycin pre-treated lungs had large tumor nodules (Figure 3B). Using this protocol, we also increased the engraftment of another human LUAD cell line into athymic mice (PC9; Figure 3C) and a murine LUAD cell line that was injected into syngeneic immunocompetent B6129SF1/J mice (Figure 3D). The early lesions observed in bleomycin pre-treated syngeneic mice were mainly grade 1 and 2 and were well differentiated, resembling tumors arising in situ from KrasG12D/+;p53-/- GEMMs33. All engrafted LUADs grew both in the proximal and distal lung (Figure 3B, Figure 3E, Figure 3G; asterisk (distal), star (proximal)). The establishment of PDXs from lung cancer biospecimens (from biopsies or surgical resections) is relatively inefficient, even when human cells are subcutaneously transplanted into severely immunodeficient animals such as the NOD scid gamma (NSG) mouse strain34. To evaluate the potential application of our method to PDXs, we also injected a highly metastatic cell sub-population of the H2030 cell line, termed H2030-BrM3 cells35, intratracheally in NSG mice that had been pre-conditioned with vehicle or bleomycin. We note that NSG mice may be more susceptible to the toxic effects of bleomycin (data not shown) and that a dose lower than 0.02 units per mouse is advisable when working with this strain. Nevertheless, our protocol also significantly increased the engraftment and tumor burden in the lungs of NSG mice (Figure 3F, Figure 3G).
We next used this protocol to test the engraftment rates of limiting amounts of cells using the H2030 cell line in athymic mice. Mice pre-treated with bleomycin were injected with 5 x 105, 5 x 104, or 5 x 103 H2030 cells. 85.7-100% of mice developed lung tumors between 7 and 10 weeks at all dilutions tested (Table 2). We conclude that engraftment of a relatively low number of tumor cells in the lung is possible if mice are pre-conditioned with bleomycin. Overall, our protocol could be tailored to several strains of mice and various LUAD cell line models to increase the engraftment rate of cancer cells into the lungs from 0-16.7% to 71.4-100% (Table 2).
Finally, to evaluate the putative kinetics of LUAD cell outgrowth and metastatic dissemination from bleomycin pre-conditioned lungs, we characterized the behavior of highly metastatic H2030-BrM3 cells using our protocol. Following orthotopic engraftment in vehicle-treated mice, tumor cell attrition was observed by bioluminescence at day 3 post-injection and lung tumor growth could not be detected thereafter or at endpoint (day 35) (Figure 4A, Figure 4B). Conversely, we detected lung tumor outgrowth in mice pre-treated with bleomycin as early as day 7 post-injection (Figure 4A). After 39-57 days of tumor expansion, the morbidity and bioluminescent intensity associated with high lung tumor burden limits the analysis of disseminated disease (data not shown). Therefore, we imaged metastasis in distant organs ex vivo at necropsy. In animals that had lung tumor burden for 57 days, small lesions could also be detected in tissues such as the brain, liver, kidney, adrenal glands, and bone hind limbs at variable frequencies (Figure 4C). Therefore, we conclude that spontaneous disseminated disease in clinically relevant sites can be generated from this orthotopic method.
Figure 1: Method to pre-condition the lungs for tumor cell engraftment. Schematic of the protocol used. Lung injury is performed by injecting bleomycin intratracheally. LUAD cells are then injected into the trachea of mice 14 days after. Mice are subsequently imaged weekly for luminescence to monitor tumor engraftment, growth and distant metastasis. Please click here to view a larger version of this figure.
Figure 2: Bleomycin causes transient lung fibrosis and ECM remodeling. Representative images of Hematoxylin-Eosin (H&E) and Masson Trichrome staining of lungs from vehicle and bleomycin treated mice 14 days post-injection (peak of fibrosis) and at 51 days post-injection (fibrosis resolution) in the absence of tumor cell injection. Scale bar = 500 µm. Please click here to view a larger version of this figure.
Figure 3: Lung engraftment of LUAD cell lines in multiple mouse models is increased in mice pre-treated with bleomycin. A) Athymic mice pre-treated with vehicle or bleomycin were injected with 5×105 H2030 cells. Relative lung tumor burden at endpoint (day 35 normalized to day 0 post-injection) as measured by bioluminescence is plotted. For each group, a mouse with median tumor burden is shown. n=6-7 mice per group. B) Representative H&E images of lungs pre-treated with vehicle or bleomycin from A at day 35. Asterisk = distal; Star = proximal. C) Athymic mice treated as in A were injected with 1×105 PC9 cells. Tumor burden and representative images are measured and shown as in A. n=7 mice per group. D–E) B6129SF1/J mice pre-treated with vehicle or bleomycin were injected with 1×105 368T1 KrasG12D/+ p53-/- (KP T-non Met) cells. Tumor burden, representative images and H&Es are measured and depicted as in A and B. n= 7 mice per group. F–G) NSG mice pre-treated with vehicle or bleomycin were injected with 1 x 105 H2030-BrM3 cells. Tumor burden, representative images and H&Es (day 9 and day 10 post-injection) are measured and shown as in A and B. n=3-6 mice per group. Scale bar = 500 µm. Inset = high magnification of tumor cell engrafted areas. Scale bar of insets = 100 µm. p values calculated by the Mann-Whitney test. Please click here to view a larger version of this figure.
Figure 4: LUAD cells engrafted in bleomycin-treated mice metastasize to multiple distant organs. A) Athymic mice pre-treated with vehicle or bleomycin were injected with 5×105 H2030-BrM3 cells. Lung tumor burden was measured weekly using bioluminescence. Dorsal total photon flux normalized to day 0 is shown. B) Tumor burden by bioluminescence of the lungs from A at endpoint (day 35 post-injection). For each group, a mouse with median tumor burden is shown. n=6-7 mice per group. C) Representative images of organs with detectable metastasis from the same mouse at necropsy (left). Table with indicated frequencies of mice with metastasis in distant organs at endpoint (day 39 to 57 post-engraftment) (right). Ex-vivo organs with visual bioluminescent signal above 1.5 x 104 photons/(sec cm2 steradian) were counted as positive metastasis. Please click here to view a larger version of this figure.
Mouse Strain | Units/mouse |
Athymic | 0.02 |
NSG | 0.02 |
B6129SF1/J | 0.005 |
Table 1: List of recommended doses of bleomycin for mouse strains tested.
Mouse Strain | Cell Line | Number of cells injected | % engraftment | p value | |
Vehicle | Bleomycin | ||||
Athymic | H2030 | 5×105 | 0% (0/7) | 100% (6/6) | 0.0006 |
Athymic | H2030 | 5×104 | n.d. | 85.7% (6/7) | n.a. |
Athymic | H2030 | 5×103 | n.d. | 85.7% (6/7) | n.a. |
Athymic | PC-9 | 1×105 | 0% (0/7) | 85.7% (6/7) | 0.0023 |
NSG | H2030-BrM3 | 1×105 | 16.7% (1/6) | 100% (3/3)* | 0.0476 |
B6129SF1/J | 368T1 | 1×105 | 0% (0/7) | 71.4% (5/7)* | 0.0105 |
Table 2: Summary table of the engraftment frequency (%) using the indicated mouse strains, cell lines and cell numbers. Engraftment was determined by in vivo luminescence imaging and confirmed by ex vivo imaging with histology of the tissues between 37 and 50 days post-injection. * Positive engraftment was confirmed by in vivo luminescence at day 10 post-injection.p values calculated by one-sided Fischer's exact test. n.a. = not applicable; n.d. = not determined.
Striking clinical parallels have been documented between lung cancer and other chronic diseases of the lung36. In particular, patients with idiopathic pulmonary fibrosis (IPF) have an increased predilection for developing lung cancer, and this association is independent of smoking history37,38. IPF is characterized by progressive destruction of lung architecture and impaired respiratory function through deposition of ECM39. Also, following surgical resection, early stage NSCLC patients with concurrent IPF have a poor outcome40. Most NSCLC tumors contain a fibrotic component at the time of diagnosis, and the extent of this stromal reaction correlates with poor prognosis18. Acute or sustained tissue damage from fibrosis may increase the incidence of tumorigenesis in multiple ways. For instance, after injury, the process of epithelial regeneration may expand the pool of progenitor cells susceptible to transformation. Alternatively, the influx and function of various immune cells may help establish an immunosuppressive microenvironment, while the activation of myofibroblasts may secrete growth factors or deposit tumor-promoting ECM. Accordingly, our method of pre-conditioning the lungs with a fibrosis inducing agent may prove to be particularly apt at studying the effects of lung cancer co-morbidities (e.g. fibrosis) and modeling NSCLC subtypes, which are rich in extracellular matrix and an inflammatory stroma41.
Overall, our method significantly enhanced the engraftment of tumor cells injected into the airways via the trachea. This observation holds true for mouse strains with varying degrees of immunocompetence. Other methods of tumor cell injection into the lungs include tail vein injection or direct injection into the lung parenchyma via the chest wall. For the specific purpose of studying lung cancer cell tumorigenicity, these methods are sub-optimal, as the former requires tumor cells to survive in circulation and extravasate before outgrowth (a process more akin to metastasis into the lungs), while the latter can cause cells to leak out into the pleural space soon after injection (data not shown). Both methods may confound loco-regional quantification of lung cancer cell outgrowth. Alternatively, our protocol ensures that seeded tumor cells are first retained in the airways surrounded by the lung stroma. Critical to the success of this protocol, are the proper intubation of the mice, tolerable dose of bleomycin, and timing of airway pre-conditioning relative to tumor cell injection. We also demonstrate that this protocol allows the cells to subsequently disseminate to other tissues, including the brain. Although the morbidity associated with lung tumor burden may still limit comprehensive characterization of distant macrometastasis, this approach may be useful to quantify and study circulating tumor cells originating from the lungs.
Despite the advantages of the method described herein, several technical variables may influence the ultimate efficiency of engraftment and monitoring of tumor progression. First, it is well documented that the fibrotic response to bleomycin in mice is strain-dependent. For instance, C57Bl/6 mice are more prone to fibrosis when compared to Balb/c mice42. Second, sensitivity to bleomycin is gender and age dependent. In our experience, females and younger mice are more prone to adverse effects over the entire protocol. This may limit tumorigenesis studies requiring direct comparisons between male and female, or young and old animals. Moreover, intratracheal instillation of mice younger than 7 weeks is technically challenging and adjustments to the catheter size may be required. Third, bleomycin is toxic and may be mutagenic due to its ability to induce DNA breaks. It is important to test the tolerable dose of bleomycin for each mouse strain prior to performing the experiments. In this proof of principle study, we provide suggested doses for male mice across various strains. Additionally, mice with dark fur are less suitable for bioluminescent imaging due to low tissue penetration and masking of the bioluminescent signal by the fur. Depilating the mice may improve image measurements, but alternative methods can also be employed such as Magnetic Resonance Imaging and fluorescent imaging using probes of long wavelength such as TdTomato or Katushka43. Finally, the potential immunogenicity of a given reporter protein should be considered when selecting a tumor detection modality for a given mouse model.
Bleomycin remodels the distal lungs by first damaging alveolar type 2 (AT2) cells which then regenerate in an attempt to repair the alveoli44. Since AT2 cells are one of the major cell types to give rise to NSCLC, pre-conditioning the lungs with bleomycin may also modify the initiation and progression of tumors arising in situ from GEMMs. Other types of injury, including influenza infection13 or naphthalene instillation14, have been shown to increase the engraftment of human and murine lung stem/progenitor cells. Notably, naphthalene damages stem/progenitor cells in the bronchio-alveolar junctions and potentially increases tumor initiation in GEMMs45. Fine-tuning the type of injury and airway niche targeted to the engrafted cell types may further optimize our method to study lung cancer allografts or xenografts of different origins. Finally, we propose that this method might be implemented for the standard generation of lung cancer PDXs. The success rate of establishing lung cancer PDXs from human biospecimens by subcutaneous transplantation in NSG mice is relatively low (30-40%)46,47. Orthotopic growth of PDXs may not only increase this success rate with limiting amounts of specimen (from human lungs), but also generate more physiologically relevant conditions to test the pharmacokinetics and pharmacodynamics of lung cancer therapeutics.
The authors have nothing to disclose.
This study was funded by grants from the National Cancer Institute (R01CA166376 and R01CA191489 to D.X. Nguyen) and the Department of Defense (W81XWH-16-1-0227 to D.X. Nguyen).
Bleomycin | Sigma | B5507-15UN | CAUTION Health hazard GHS08 |
Exel Catheter 24G | Fisher | 1484121 | Remove needle. For intratracheal injection |
Ketamine (Ketaset inl 100 mg/mL C3N 10 mL) | Butler Schein | 56344 | To anesthetize mice |
Xylazine | Butler Schein | 33198 | To anesthetize mice |
Ketoprofen, 5,000 mg | Cayman Chemical | 10006661 | Analgesic |
Puralube Veterinary Ophthalmic Ointment | BUTLER ANIMAL HEALTH COMPANY LLC | 8897 | To prevent eye dryness while under anesthesia |
D-Luciferin powder | Perkin Elmer Health Sciences Inc | 122799 | For luminescent imaging. Reconstitute powder with PBS for a working concentration of 15mg/mL. Protect from Light |
Rodent Intubation stand | Braintree Scientific | RIS-100 | Recommended stand for intratracheal injection |
MI-150 ILLUMINATOR 150W MI-150 | DOLAN-JENNER INDUSTRIES | MI-150 / EEG2823M | To illuminate and visualize trachea |
Graefe Forceps, 2.75 (7 cm) long serrat | Roboz | RS-5111 | For intratracheal injection |
Syringe Luer-Lok Sterile 5ml | BD / Fisher | 309646 | |
Satiny Smooth by Conair Dual Foil Wet/Dry Rechargeable Shaver | Conair | – | To shave mice |
Bonn Scissors, 3.5" straight 15 mm sharp/sharp sure cut blades | Roboz | RS-5840SC | |
15 mL conical tube | BD / Fisher | 352097 | |
1.5 mL centrifuge tubes | USA SCIENTIFIC INC | 1615-5500 | |
Vial Scintillation 7 mL Borosilicate Glass GPI | Fisher | 701350 | |
Filter pipette tips (200 μL) | USA SCIENTIFIC INC | 1120-8710 | |
Phosphate Buffered Saline | Life Technologies | 14190-144 | |
0.25% Trypsin-EDTA | Life Technologies | 25200-056 | |
DMEM high glucose | Life Technologies | 11965-092 | |
RPMI Medium 1640 | Life Technologies | 11875-093 | |
Fetal bovine serum USDA | Life Technologies | 10437-028 | |
Penicillin-Streptomycin | Life Technologies | 15140-122 | |
Amphotericin B | Sigma | A2942-20ML | |
Trypan Blue Stain 0.4% | Life Technologies | 15250-061 | |
Countess Automated Cell Counter | Life Technologies | AMQAX1000 | |
Flask T/C 75cm sq canted neck, blue cap | Fisher / Corning | 353135 | |
IVIS Spectrum Xenogen Bioluminiscence | Perkin Elmer Health Sciences Inc | 124262 | For in vivo bioluminescence imaging |
Living image software | Perkin Elmer Health Sciences Inc | 128113 | For in vivo bioluminescence analysis |
XGI-8 Gas Anesthesia System | Perkin Elmer Health Sciences Inc | 118918 | For Isoflurane anesthesia |
BD Ultra-Fine II Short Needle Insulin Syringe 1 cc. 31 G x 8 mm (5/16 in) | BD / Fisher | BD328418 | For retro-orbital luciferin injection |
Syringe 1ml | BD / Fisher | 14-823-434 | For intraperitoneal injections |
26 G x 1/2 in. needle | BD / Fisher | 305111 | For intraperitoneal injections |
4% Paraformaldehyde | VWR | 43368-9M | CAUTION Health hazard GHS07, GHS08. For fixing tissue |
Pipet-Lite Pipette, Unv. SL-200XLS+ | METTLER-TOLEDO INTERNATIONAL | 17014411 | |
Mayer's Hematoxylin | ELECTRON MICROSCOPY SCIENCES | 517-28-2 | |
Eosin Y stain 0.25% (w/v) in 57% | Fisher | 67-63-0 | |
Masson Trichrome Stain Kit | IMEB Inc | K7228 | For masson trichrome stain to visualize collagen |
Superfrost plus glass slides | Fisher | 1255015 | |
6 well plate | Corning | C3516 | |
Universal Mycoplasma Detection Kit | ATCC | 30-1012K | |
OCT Embedding compound | ELECTRON MICROSCOPY SCIENCES | 62550-12 | For embedding tissue for frozen sections |
Leica CM3050 S Research Cryostat | Leica | CM3050 S | To section tissue for staining analysis |
Keyence All-in One Fluorescence Microscope | Keyence | BZ-X700 | |
ImageJ | US National Institutes of Health | IJ1.46 | http://rsbweb.nih.gov/ij/ download.html |
Prism 7.0 for Mac OS X | GraphPad Software, Inc. | – | |
Athymic (Crl:NU(NCr)-Foxn1nu) mice | Charles River | NIH-553 | |
NSG (NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ) mice | Jackson Laboratories | 5557 | |
B6129SF1/J mice | Jackson Laboratories | 101043 | |
NIH-H2030 cells | ATCC | CRL-5914 | |
368T1 | generously provided by Monte Winslow (Standford University) | – | |
PC9 cells | Nguyen DX et al. Cell. 2009;138:51–62 | – | |
H2030 BrM3 cells | Nguyen DX et al. Cell. 2009;138:51–62 | – |