Julien Sage, PhD
Experimental model systems are an indispensable part of research into SCLC. At the International Association for the Study of Lung Cancer’s 2021 Hot Topic Meeting: Small Cell Lung Cancer, keynote speaker Julien Sage, PhD, discussed the importance of these experimental models, specifically mouse models, in moving the field forward.
Dr. Sage, who is the Elaine and John Chambers Professor of Pediatric Cancer and Professor of Genetics at Stanford University, opened his presentation with a reminder to the audience that survival for patients with SCLC has not improved much during the past 3 to 4 decades and remains a significant clinical issue.
“One way to approach this is to gain a better understanding of the biology of the disease,” Dr. Sage said. “If we understand the biology better, we might be able to identify new ways to target SCLC.”
Another important goal is to develop better preclinical models of SCLC. Better models allow for more precise testing of new targeted therapies or other new ideas that could eventually improve the treatment of patients with SCLC.
Dr. Sage acknowledged that there are a large number of investigators exploring the biology of SCLC and experimenting with improving preclinical models. He said the ideal scenario would be to study new therapeutics in primary human tumors, of course. However access to primary tumor specimens is not always easy and clinical trials are often challenging to implement.
Researchers also have cell lines that can be used for preclinical experiments. For example, there is a lot of interest in using patient-derived xenografts and circulating tumor cell-derived xenografts (see related article on the keynote presentation by Dr. Caroline Dive). There is also a lot of excitement around tumor organoids for SCLC, Dr. Sage said.
However, during his presentation, Dr. Sage focused on the use of genetically engineered mouse models in SCLC. These mouse models can be used to study all stages of disease, from early lesions to metastatic disease.
“In mice, SCLC initiation is controlled, and tumor growth is measurable, including using reporters that allow us to monitor tumors in vivo,” Dr. Sage said.
In addition, SCLC tumors growing in genetically engineered mouse models functionally interact with immune cells, which can be critical in the development of immunotherapies.
Available Models
The first reported genetically engineered mouse model was developed by the group of Dr. Anton Berns by inactivating the Trp53 and Rb1 tumor suppressor genes in the lungs of young adult mice. These two genes are nearly universally inactivated in human tumors and are thought to be key gatekeepers to prevent SCLC initiation. Dr. Sage’s lab later added another cancer gene to the mix, the Rb1-related gene p130 (also known as Rbl2). This model with inactivation of three tumor suppressors was referred to as the triple knockout model (TKO). It’s also known as RPR2. In this context, loss of p130 initiates more tumors and accelerates SCLC development compared to the double knockout model. A third mouse model added later is the RPM model developed by the lab of Dr. Trudy Oliver, in which high expression of the Myc oncogene is combined with loss of Rb1 and Trp53. Although there are numerous other mouse models of SCLC, the RP, RPR2, and RPM models are the three models with the most data.
Importantly, the histology of the tumors that develop in these mice is quite similar to human SCLC, Dr. Sage said. The tumors metastasize to a number of sites outside the lungs and, overall, respond similarly to both chemotherapy and immunotherapy. Interestingly, while the mouse tumors are initiated with just two to three genetic lesions, they gain new alterations during tumor progression and these new alterations are similar to recurrent alterations detected in human tumors, including mutations in genes such as Notch or activations of oncogenes such as L-Myc.
Dr. Sage noted that there are, however, limitations to mouse models. First, he said, these models are slow. It takes time to breed mice and grow tumors. The process can take 2 to 4 years depending on the model and the type of analysis for each new gene of interest.
Additionally, unlike human SCLC, these mice often die from tumor burden in the lung rather than metastases. Moreover, brain metastases occur frequently in human SCLC, but it has been difficult to observe brain metastases in the brains of the mouse models, Dr. Sage noted.
Finally, patients with SCLC are almost always smokers and have many mutations as a result of carcinogens present in tobacco smoke. The mouse models have fewer alterations in their genome, which may change inflammation levels and the activity of certain immune cells.
Model Improvements
Because of these limitations, work is being done to try to improve mouse models of SCLC. One of the first ways being researched is the use of allografts of SCLC. In this approach, single tumors are isolated from mice and propagated in other mice from the same genetic background, with an intact immune system. For example, some experiments have been done successfully using these allograft models to test T-cell checkpoint blockade.
Other ongoing work is being done to model brain metastases with allografts, Dr. Sage said. He and members of his lab are testing the hypothesis that neuronal characteristics of SCLC cells allow them to functionally interact with other cells in the brain microenvironment. When they looked at this interaction in allograft models directly injected into the brain of mice,1 they noticed a strong and close interaction between cancer cells and GFAP-positive astrocytes. The cancer cells reactivate the astrocytes, Dr. Sage said, and the astrocytes, when in close proximity to cancer cells, produce pro-survival factors to promote growth of cancer cells.
They also identified Reelin, a factor involved normally in brain development, as a factor secreted by SCLC cells to recruit astrocytes, indicating that neuronal Reelin produced by SCLC is critical for the growth of SCLC brain metastases.
Another way to help enhance SCLC mouse models would be different viral platforms to initiate tumors. In particular, Dr. Sage discussed the use of lentiviral vectors to express the Cre recombinase to inactivate the Trp53 and Rb1 tumor suppressor genes (flanked by lox sites, the targets of Cre). This approach remains inefficient, as recently evidenced by a study from the Rudin lab.2 Lentiviral vectors could also allow for the stable, long-term expression of other molecules, including sgRNA molecules to knockout genes of interest or cDNAs to overexpress other genes of interest.
“We would also be interested in using lentiviruses to induce and track tumors,” Dr. Sage said.
The Sage lab is developing protocols to enhance the activity of lentiviruses in mouse models of SCLC, including combining lentiviruses with molecules inducing injury in the lungs to possibly activate the expansion of cells that may serve as cell-of-origin for SCLC.
References
- 1. Qu F, Cao S, Michno W, et al. Neuronal mimicry generates an ecosystem critical for brain metastatic growth of SCLC. Biorxiv.org. August 10, 2021. https://www.biorxiv.org/content/10.1101/2021.08.10.455426v1. Accessed December 14, 2021.
- 2. Rlf-Mycl Gene Fusion Drives Tumorigenesis and Metastasis in a Mouse Model of Small Cell Lung Cancer. Cancer Discov. 2021;11(12):2945.
