Research Roundup: March 5, 2021

Welcome to the March 5, 2021 installment of Research Roundup, a recurring snapshot of recent studies published by scientists at the Broad Institute and their collaborators.

Super-sized single-cell study of SARS-CoV-2 susceptibility

A team including leaders of the Human Cell Atlas’s Lung Biological Network and led by trainees from the Broad Institute and Helmholtz Center Munich has completed one of the first meta-analyses of single-cell studies examining biological risk factors for COVID-19. They found that healthy cells in diverse tissues express SARS-CoV-2 entry factors, which are more abundant in certain cells from people at risk of severe COVID due to age, gender, or smoking status. The work suggests molecular roots for COVID risk factors and offers a rich resource for studies of the virus and potential therapies. Read more in Nature Medicine or on the Broad website.

The immunology of allergy 

Mast cells (MCs), a type of immune cells, are associated with type 2 (T2) allergic inflammatory diseases of the airway, including asthma and chronic rhinosinusitis with nasal polyposis (CRSwNP). However, there is a limited understanding of the molecular mechanisms through which MCs impact T2 disease. Daniel Dwyer (Harvard Medical School), associate member Jose Ordovas-Montanes at the Klarman Cell Observatory (KCO), institute member Alex Shalek, Nora Barrett (HMS), Joshua Boyce (HMS), and colleagues conducted a detailed study of human MC hyperplasia (unusually high growth of cells) in CRSwNP using flow cytometry and single-cell RNA sequencing. In Science Immunology, the researchers report that localized proliferation of MCs rich in specific transmembrane glycoproteins such as CD38 and CD117 contributes to T2 disease.

Applying deep learning to computational pathology

Deep learning is revolutionizing the analysis of medical images, but studying gigapixel whole-slide images (WSIs) remains a challenge due to their complexity. In Nature Biomedical Engineering, a team led by Ming Yu and Cancer Program associate member Faisal Mahmood reports a more efficient method of processing WSIs and mining them for relevant insights. The researchers applied this approach, called CLAM (clustering-constrained-attention multiple-instance learning), to differentiate subtypes of renal cell carcinomas and non-small-cell lung cancers, and to detect lymph node metastasis. The team aims for CLAM to offer a new way to solve clinical and research problems using WSIs of routine histology specimens. 

Essential genes vary depending on nutrient availability

CRISPR-based screens have identified genes that are essential to many cancer cell lines, inspiring research on cancer therapies that target those genes. However, most screens are performed in cell culture media that doesn’t reflect conditions found in the human body. In Cell Metabolism, Nicholas Rossiter (UW Madison), David Sabatini, Jason Cantor (UW Madison), and colleagues performed a handful of CRISPR screens in normal culture media versus media that mimics human plasma. They found that the essential genes differed depending on the media used, and that these differences depended on the nutrient availability in each media type.

APOE4 disrupts lipid metabolism

APOE4 is a strong genetic risk factor for many diseases, including late-onset Alzheimer’s disease. Cell Circuits Program associate member Li-Huei Tsai (MIT) and colleagues showed that cultured human glia with an APOE4 genotype accumulated unsaturated triglycerides, resulting in a lipid imbalance. Promoting phospholipid synthesis through choline supplementation of culture medium restored a normal lipid state in APOE4-expressing yeast cells. The authors found that choline supplementation also restored lipid homeostasis in human APOE4 astrocytes derived from iPS cells. These findings, in Science Translational Medicine, suggest that modulating glial metabolism could be a therapeutic approach for APOE4-associated diseases.

Snapshots of resistance

Metastatic prostate cancer is nearly universally treatment resistant and lethal. To better understand the roots of its resistance — and identify opportunities for counteracting it — Meng Xiao He, Mary Ellen Taplin (DFCI), Cancer Program associate member Eliezer Van Allen, and collaborators (including many in the KCO, including Mike Cuoco and core institute member Aviv Regev [now at Genentech]) analyzed single cell gene expression profiles of tumor and immune cells from 14 advanced prostate cancer patients. They found expression in tumor cells of more than one form of the cancer-driving androgen receptor at a time, resistance-associated cell states that could be vulnerable to TGF-β blockade, and immune cell signatures suggesting potential immunotherapy strategies. Learn more in Nature Medicine.

New rate estimates for structural variants

Geneticists have a good handle on the rate of new (de novo) point mutations in the human germline, but not for that of de novo structural mutations (dnSVs). To help do so, Harrison Brand, Harold Wang, Xuefang Zhao, and institute member Michael Talkowski in the Program in Medical and Population Genetics and colleagues profiled germline mutations in more than 18,600 people. In the American Journal of Human Genetics, they report that dnSVs arise far less frequently (≥0.160 mutations per person in general, 0.260 in autism) than point mutations (~70 per person), that nearly three-fourths of dnSVs arise in parents' eggs or sperm, and that dnSV rate is not correlated with parental age.

To learn more about research conducted at the Broad, visit broadinstitute.org/publications, and keep an eye on broadinstitute.org/news.