One cancer, two causes, many immune responses


                               Anthony Cillo

“Using CRC gives me more time to explore, to write code, to develop pipelines – and to take risks and make mistakes that don’t cost tens of thousands of dollars. Without CRC, we would have had to spend a ton of money and a lot of the valuable discovery in the paper would not have been possible. CRC is an amazing resource at Pitt.”

Head and neck cancer in men has historically been attributed to exposure to carcinogens like smoking. But increasingly the Human Papillomavirus, which contributes to cervical cancer in women, is identified as the cause of head and neck cancer in men (technically, Head and Neck Squamous Cell Carcinoma, or HNSCC). HNSCC provokes very different immune responses in different patients – survival rates are low for all patients, but those with viral cancers have better overall survival rates. The same malignancy arises in the same location through different mechanisms, triggering different immune responses with different results. What could that mean for treating the disease?

Anthony Cillo is a post-doc in the laboratory of Dario Vignali, vice chair of the Department of Immunology and leader of Pitt’s Cancer Immunology Program, one of a number of Immunology researchers who collaborate with the Center for Research Computing. Cillo explored the questions of the variability of immune responses in HNSCC as lead author of a study published in January in the journal Immunity – “Immune landscape of viral- and carcinogen- driven head and neck cancer.” Cillo and collaborators used CRC’s advanced resources in conjunction with single-cell RNA sequencing technology to examine immune cells in blood and tumors from HNSCC patients.

The Immunity paper demonstrated that the difference between the anti-tumor immune response in the viral and environmental cancers can be credited to a specific population of T-cells, called follicular helper T-cells, which may be key to designing improved and expanded immunotherapies.

“Immunotherapy has great potential,” explains Cillo. “But right now, only a small percentage of patients respond. It is becoming more standard as a second line therapy after traditional radiation, surgery, and chemotherapy, and it is in clinical trials as a frontline therapy. The question we try to answer in the study is why some patients respond, and some do not, and how can we develop analytical tools to help understand those differences and identify targets for treatment.”

The work began with tissue samples taken at Hillman Cancer Center, often coming to Cillo only an hour after they have been removed. Using samples of patients’ blood and HNSCC tumors, Cillo chemically separated live CD45+ cells – signaling molecules whose proteins play a role in regulating immune cell responses – from other cells. He sequenced a total of 131,224 individual cells to find genomic profiles of thousands of single cells in order to identify subpopulations with different properties.

Cillo collaborated with CRC consultant Fangping Mu in using CRC’s High Throughput Cluster (HTC) for computationally intensive processing of the next-generation sequencing data, producing around 50 gigabytes of data, a process that consumed roughly 60,000 computing hours. He then used HTC to analyze the processed single-cell RNAseq data by clustering subpopulations, working with datasets up to several hundred thousand cells, with roughly 30,000 genes.

Cillo builds his own pipelines – workflows of analysis programs – using HTC to develop new tools and optimize existing tools for computational analysis of single-cell RNAseq data using an algorithm that draws on multiple cores for critical steps, speeding up the analysis compared to standard existing algorithms.

Cillo describes CRC as invaluable for allowing him the freedom to build custom tools.

“Using CRC gives me more time to explore, to write code, to develop pipelines – and to take risks and make mistakes that don’t cost tens of thousands of dollars. Without CRC, we would have had to spend a ton of money and a lot of the valuable discovery in the paper would not have been possible. CRC is an amazing resource at Pitt.”

 

Contact:
Brian Connelly
Center for Research Computing
bgc14@pitt.edu