NOTE: HTA data is currently organized by the original publication. As these data are incorporated into multi-functional Atlases, search and other means of accessing data will be added.

The spatial landscape of progression and immunoediting in primary melanoma at single cell resolution

Ajit J. Nirmal, Zoltan Maliga, Tuulia Vallius, Brian Quattrochi, Alyce A. Chen, Connor A. Jacobson, Roxanne J. Pelletier, Clarence Yapp, Raquel Arias-Camison, Yu-An Chen, Christine G. Lian, George F. Murphy, Sandro Santagata, Peter K. Sorger

Cancer Discovery. 2022 Jun; 12(6): 1518-1541. PMID: 35404441

Cutaneous melanoma is a highly immunogenic malignancy, surgically curable at early stages, but life- threatening when metastatic. Here we integrate high-plex imaging, 3D high-resolution microscopy, and spatially-resolved micro-region transcriptomics to study immune evasion and immunoediting in primary melanoma. We find that recurrent cellular neighborhoods involving tumor, immune, and stromal cells change significantly along a progression axis involving precursor states, melanoma in situ, and invasive tumor. Hallmarks of immunosuppression are already detectable in precursor regions. When tumors become locally invasive, a consolidated and spatially restricted suppressive environment forms along the tumor-stromal boundary. This environment is established by cytokine gradients that promote expression of MHC-II and IDO1, and by PD1-PDL1 mediated cell contacts involving macrophages, dendritic cells, and T cells. A few millimeters away, cytotoxic T cells synapse with melanoma cells in fields of tumor regression. Thus, invasion and immunoediting can co-exist within a few millimeters of each other in a single specimen.

Publication | bioRxiv

HTA MEL Atlas 1

The HTA MEL Atlas 1 dataset contains images and other data being used for construction of an atlas of human melanoma under the auspices of the Human Tumor Atlas Network. Advanced solid cancers are complex assemblies of tumor, immune, and stromal cells that invade adjacent tissue and spread to distant sites. We use highly multiplexed tissue imaging, spatial statistics, and machine learning to identify cell types and states underlying morphological features of known diagnostic and prognostic significance in melanoma. This includes the tumor invasive margin, where tumor, normal, and immune cells compete and were diverse immunosuppressive environments are found.

Contents

Data Explorations

Data Explorations are like museum guides and exploit the digital docents in MINERVA to guide readers through the complexities of a large image dataset via a series of narrated stories and waypoints.

The images in Nirmal et al. (2021) comprise a ~2.3 TB dataset with some images as large as 1 gigapixel. With MINERVA, users can pan around and magnify areas of an image and switch between channels. MINERVA does not require the installation of any software and is therefore secure; browsing is also anonymous. Users interested in the tool are welcome to explore the documentation, the software publication, and a description of digital docents in general.

Data image
Introduction to the MEL Atlas (MEL 1 Abstract)
Data image
Deep Exploration of a Primary Melanoma (MEL 1 Full Story)

Data overviews

Data Overviews provide access to minimally processed Level 2 images with no annotation or quality control. Click any of the following thumbnail images for an interactive view of the full-resolution images.

Data image
MEL01-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL01-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL01-1-3 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL01-1-4 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL01-1-5 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL01-1-6 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL01-2-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL01-2-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL01-3-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL01-3-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL02-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL02-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL03-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL03-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL04-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL04-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL05-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL05-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL06-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL06-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL07-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL07-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL08-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL08-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL09-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL09-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL10-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL10-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL11-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL11-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL12-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL12-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL13-1-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL13-1-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL13-2-0 - H&E - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021
Data image
MEL13-2-1 - overview - Nirmal, Maliga, Vallius, Sorger et al., 2021

Primary Data Access

About the data files

The primary data represents minimally processed (Level 2) image data from either 2D whole slide or 3D optically-sectioned imaging relevant to HMS HTAN Center Melanoma Atlas 1. Whole slide images are segmented, quantified, and subjected to additional quality control to generate final data. Whole slide scans are saved as OME-TIFF tiled pyramid images whereas 3D datasets are three-dimensional TIFF files. Spatial feature tables are zipped CSV files.

Data were collected using cyclic immunofluorescence (CyCIF), as described in https://dx.doi.org/10.17504/protocols.io.bjiukkew, or brightfield imaging of hematoxylin and eosin (H&E) stained slides. Each file corresponds to a multiplexed image mosaic for FFPE tissue sections 5 microns thick with sample diameter extending up to ~2.9 cm. Each whole slide image is assembled from a series of successive image tiles stitched together (832 x 732 µm tiles; up to 990 tiles/image) and flat-field corrected for illumination using MCMICRO software to generate ~0.2 to ~1.5 gigapixel images. Tiles were collected for each CyCIF round (up to 15 in total depending of the antibody panel used) and these are combined in the mosaic image to generate a composite with 32 to 60 channels. Each 3D image (110 x 110 x 16 µm) requires 3D image registration to assemble all 7 CyCIF rounds.

Whole slide CyCIF images were collected on a RareCyte Inc. CyteFinder slide scanning fluorescence microscope using a 20x/0.75 NA objective and sampled at 650 nm/pixel. 3D high resolution CyCIF images were acquired on a GE Deltavision Elite equipped with a 60x/1.42 NA oil immersion objective lens and sampled at 108 nm/pixel in X & Y, and 200 nm steps in Z axis. H&E images were collected on an Olympus VS120 microscope using a 20x/0.75 NA objective and sampled at 350 nm/pixel.

The files MEL01-1-* to MEL01-3-* derive from a 62-year old male (MEL1), who had a stage IIC primary melanoma with NF1 (c.1008G>A and c.4006C>T) mutation. The tumor had invaded into reticular dermis and was surgically removed.

The files MEL02-1-* to MEL13-2-* derive from 12 additional patients from the Brigham and Women’s Hospital. Additional information on antibodies and specimen are available at https://labsyspharm.github.io/HTA-MELATLAS-1/.

Image files ending in -0-ROI* are H&E images. All others are CyCIF images.

Download the primary data

Download the primary data

The full dataset is available through Amazon Web Services S3 using a “requester pays” model. AWS charges $0.10/GB for downloading all or part of the data. The person downloading the data must have an AWS account and must email tissue-atlas@hms.harvard.edu with the AWS account’s AWS account ID and canonical user ID which may be found as detailed here: https://docs.aws.amazon.com/general/latest/gr/acct-identifiers.html#FindingYourAccountIdentifiers. We must receive both the account ID and canonical user ID in order to grant access to the S3 bucket containing the full dataset. After access is granted the images and metadata will be available in the bucket at the following location:

s3://hta-melatlas-1/data/

To browse and download the data use either a graphical file transfer application that supports S3 such as CyberDuck, or the AWS CLI tools. A graphical tool may be more convenient but the CLI tools will likely offer higher download speeds. Please refer to the documentation for your chosen tool on how to sign in and enable access to requester-pays buckets. There is unfortunately no web-browser-based mechanism for accessing a requester-pays bucket. Keep in mind the download costs, which will run over $200 for downloading one copy of the entire dataset. For users who wish to perform processing within AWS to avoid transfer charges, note that the bucket is located in the us-east-1 region so any other resources must be instantiated in this same region.

The Laboratory of Systems Pharmacology at Harvard Medical School will commit to making the full dataset available through S3 while the Human Tumor Atlas Network’s Data Coordination Center exploring options to host the data.

The following table contains summary biospecimen and file metadata.

Whole-slide images (OME-TIFF) and spatial feature tables (CSV)

FIELD1 Level 2 WSI (OME-TIFF) File size (WSI) Spatial feature tables (CSV) File size (feature table)
MEL01-1-0-ROI1 MEL01-1-0-HE-ROI1.ome.tif 15 GB    
MEL01-1-1 MEL01-1-1.ome.tif 195 GB    
MEL01-1-3 MEL01-1-3.ome.tif 98 GB MEL01-1-3-features.zip 335 MB
MEL01-1-4 MEL01-1-4.ome.tif 66 GB MEL01-1-4-features.zip 200 MB
MEL01-1-5 MEL01-1-5.ome.tif 129 GB MEL01-1-5-features.zip 280 MB
MEL01-1-6 MEL01-1-6.ome.tif 174 GB MEL01-1-6-features.zip 290 MB
MEL01-2-0-ROI1 MEL01-2-0-HE-ROI1.ome.tif 13 GB    
MEL01-2-1 MEL01-2-1.ome.tif 163 GB MEL01-2-1-features.zip 82 MB
MEL01-3-0-ROI1 MEL01-3-0-HE-ROI1.ome.tif 17 GB    
MEL01-3-1 MEL01-3-1.ome.tif 230 GB MEL01-3-1-features.zip 280 MB
MEL02-1-0-ROI1 MEL02-1-0-HE-ROI1.ome.tif 0.5 GB    
MEL02-1-0-ROI2 MEL02-1-0-HE-ROI2.ome.tif 0.4 GB    
MEL02-1-1 MEL02-1-1.ome.tif 19 GB MEL02-1-1-features.zip 16 MB
MEL03-1-0-ROI1 MEL03-1-0-HE-ROI1.ome.tif 6 GB    
MEL03-1-0-ROI2 MEL03-1-0-HE-ROI2.ome.tif 5 GB    
MEL03-1-1 MEL03-1-1.ome.tif 189 GB MEL03-1-1-features.zip 160 MB
MEL04-1-0-ROI1 MEL04-1-0-HE-ROI1.ome.tif 5 GB    
MEL04-1-0-ROI2 MEL04-1-0-HE-ROI2.ome.tif 4 GB    
MEL04-1-1 MEL04-1-1.ome.tif 179 GB MEL04-1-1-features.zip 61 MB
MEL05-1-0-ROI1 MEL05-1-0-HE-ROI1.ome.tif 0.4 GB    
MEL05-1-0-ROI2 MEL05-1-0-HE-ROI2.ome.tif 0.6 GB    
MEL05-1-1 MEL05-1-1.ome.tif 34 GB MEL05-1-1-features.zip 17 MB
MEL06-1-0-ROI1 MEL06-1-0-HE-ROI1.ome.tif 17 GB    
MEL06-1-1 MEL06-1-1.ome.tif 220 GB MEL06-1-1-features.zip 299 MB
MEL07-1-0-ROI1 MEL07-1-0-HE-ROI1.ome.tif 4 GB    
MEL07-1-0-ROI2 MEL07-1-0-HE-ROI2.ome.tif 3 GB    
MEL07-1-1 MEL07-1-1.ome.tif 122 GB MEL07-1-1-features.zip 41 MB
MEL08-1-0-ROI1 MEL08-1-0-HE-ROI1.ome.tif 1 GB    
MEL08-1-0-ROI2 MEL08-1-0-HE-ROI2.ome.tif 0.9 GB    
MEL08-1-1 MEL08-1-1.ome.tif 40 GB MEL08-1-1-features.zip 29 MB
MEL09-1-0-ROI1 MEL09-1-0-HE-ROI1.ome.tif 0.8 GB    
MEL09-1-0-ROI2 MEL09-1-0-HE-ROI2.ome.tif 0.6 GB    
MEL09-1-1 MEL09-1-1.ome.tif 32 GB MEL09-1-1-features.zip 6 MB
MEL10-1-0-ROI1 MEL10-1-0-HE-ROI1.ome.tif 0.5 GB    
MEL10-1-0-ROI2 MEL10-1-0-HE-ROI2.ome.tif 0.6 GB    
MEL10-1-1 MEL10-1-1.ome.tif 34 GB MEL10-1-1-features.zip 12 MB
MEL11-1-0-ROI1 MEL11-1-0-HE-ROI1.ome.tif 2 GB    
MEL11-1-0-ROI2 MEL11-1-0-HE-ROI2.ome.tif 2 GB    
MEL11-1-1 MEL11-1-1.ome.tif 79 GB MEL11-1-1-features.zip 21 MB
MEL12-1-0-ROI1 MEL12-1-0-HE-ROI1.ome.tif 2 GB    
MEL12-1-0-ROI2 MEL12-1-0-HE-ROI2.ome.tif 1 GB    
MEL12-1-1 MEL12-1-1.ome.tif 56 GB MEL12-1-1-features.zip 14 MB
MEL13-1-0-ROI1 MEL13-1-0-HE-ROI1.ome.tif 4 GB    
MEL13-1-0-ROI2 MEL13-1-0-HE-ROI2.ome.tif 3 GB    
MEL13-1-1 MEL13-1-1.ome.tif 129 GB MEL13-1-1-features.zip 46 MB
MEL13-2-0-ROI1 MEL13-2-0-HE-ROI1.ome.tif 4 GB    
MEL13-2-0-ROI2 MEL13-2-0-HE-ROI2.ome.tif 3 GB    
MEL13-2-1 MEL13-2-1.ome.tif 121 GB MEL13-2-1-features.zip 36 MB

3D high-resolution images

FIELD1 3D image file File size
MEL01-1-4-3D-1 MEL01-1-4-3D-1.ome.tif 5 GB
MEL01-1-4-3D-10 MEL01-1-4-3D-10.ome.tif 5 GB
MEL01-1-4-3D-11 MEL01-1-4-3D-11.ome.tif 5 GB
MEL01-1-4-3D-12 MEL01-1-4-3D-12.ome.tif 4 GB
MEL01-1-4-3D-13 MEL01-1-4-3D-13.ome.tif 5 GB
MEL01-1-4-3D-14 MEL01-1-4-3D-14.ome.tif 5 GB
MEL01-1-4-3D-15 MEL01-1-4-3D-15.ome.tif 5 GB
MEL01-1-4-3D-16 MEL01-1-4-3D-16.ome.tif 5 GB
MEL01-1-4-3D-17 MEL01-1-4-3D-17.ome.tif 5 GB
MEL01-1-4-3D-18 MEL01-1-4-3D-18.ome.tif 5 GB
MEL01-1-4-3D-19 MEL01-1-4-3D-19.ome.tif 8 GB
MEL01-1-4-3D-2 MEL01-1-4-3D-2.ome.tif 5 GB
MEL01-1-4-3D-20 MEL01-1-4-3D-20.ome.tif 5 GB
MEL01-1-4-3D-21 MEL01-1-4-3D-21.ome.tif 5 GB
MEL01-1-4-3D-22 MEL01-1-4-3D-22.ome.tif 5 GB
MEL01-1-4-3D-23 MEL01-1-4-3D-23.ome.tif 4 GB
MEL01-1-4-3D-24 MEL01-1-4-3D-24.ome.tif 5 GB
MEL01-1-4-3D-25 MEL01-1-4-3D-25.ome.tif 5 GB
MEL01-1-4-3D-26 MEL01-1-4-3D-26.ome.tif 5 GB
MEL01-1-4-3D-27 MEL01-1-4-3D-27.ome.tif 5 GB
MEL01-1-4-3D-28 MEL01-1-4-3D-28.ome.tif 4 GB
MEL01-1-4-3D-29 MEL01-1-4-3D-29.ome.tif 4 GB
MEL01-1-4-3D-3 MEL01-1-4-3D-3.ome.tif 5 GB
MEL01-1-4-3D-30 MEL01-1-4-3D-30.ome.tif 5 GB
MEL01-1-4-3D-31 MEL01-1-4-3D-31.ome.tif 4 GB
MEL01-1-4-3D-32 MEL01-1-4-3D-32.ome.tif 5 GB
MEL01-1-4-3D-33 MEL01-1-4-3D-33.ome.tif 4 GB
MEL01-1-4-3D-34 MEL01-1-4-3D-34.ome.tif 4 GB
MEL01-1-4-3D-35 MEL01-1-4-3D-35.ome.tif 4 GB
MEL01-1-4-3D-36 MEL01-1-4-3D-36.ome.tif 4 GB
MEL01-1-4-3D-37 MEL01-1-4-3D-37.ome.tif 4 GB
MEL01-1-4-3D-38 MEL01-1-4-3D-38.ome.tif 4 GB
MEL01-1-4-3D-39 MEL01-1-4-3D-39.ome.tif 4 GB
MEL01-1-4-3D-4 MEL01-1-4-3D-4.ome.tif 5 GB
MEL01-1-4-3D-40 MEL01-1-4-3D-40.ome.tif 6 GB
MEL01-1-4-3D-41 MEL01-1-4-3D-41.ome.tif 4 GB
MEL01-1-4-3D-42 MEL01-1-4-3D-42.ome.tif 4 GB
MEL01-1-4-3D-5 MEL01-1-4-3D-5.ome.tif 5 GB
MEL01-1-4-3D-6 MEL01-1-4-3D-6.ome.tif 5 GB
MEL01-1-4-3D-7 MEL01-1-4-3D-7.ome.tif 5 GB
MEL01-1-4-3D-8 MEL01-1-4-3D-8.ome.tif 5 GB
MEL01-1-4-3D-9 MEL01-1-4-3D-9.ome.tif 4 GB