MYC-driven synthesis of Siglec ligands is a glycoimmune checkpoint

Significance MYC is one of the most frequently dysregulated oncogenes in human cancer. We discover that MYC causally regulates glycosylation on the surface of cancer cells, which in turn facilitates immune evasion. Using a conditional transgenic model of MYC-induced tumorigenesis, we find that MYC drives the display of a particular glycan known as disialyl-T on tumor cells. Remarkably, the disialyl-T glycan engages specific Siglec receptors on myeloid cells to inhibit the anticancer immune response, thereby promoting tumor growth in vivo. These data identify a signature of malignant glycosylation on MYC-driven cancers that suggests potential targets for immunotherapy.

This PDF file includes: 25 Supplemental Materials and Methods Figs. S1 to S24 Supplementary Table Legends  SI References  30 Other supplementary materials for this manuscript include the following: Tables S1 to S6 35

Supplemental Materials and Methods
Statistics. Measurements were made on independent samples unless otherwise noted.
Nonparametric tests were used when data could not be assumed to follow a normal distribution or verified to do so by plotting. Test statistics and effect sizes were calculated in R(1) and Prism 5 (GraphPad). General data wrangling was performed with the tidyverse(2) package in R.
Reagents and general synthetic methods. Unless stated otherwise, reactions were conducted in oven-dried glassware under an atmosphere of nitrogen using anhydrous solvents.
Tetrahydrofuran (THF) and dichloromethane (DCM) were purified by first purging with dry 10 nitrogen, followed by passage through columns of activated alumina. Deionized water was purified to 18 MΩ-cm using a Millipore Milli-Q Biocel A10 purification unit. All commercially obtained reagents were used as received without further purification unless otherwise stated.
Column chromatography was done with Biotage SNAP KP-Sil (FSK0-1107) and an Isolera Prime ACI automated fraction collector from Biotage. NMR spectra were obtained on Varian spectrometers at room temperature at the Stanford Department of Chemistry NMR Facility. 20 Retro-and lentiviral transduction. Phoenix cells containing the plasmids pHIT60 (gag-pol) and pHIT123 (ecotropic envelope) (provided by G.P. Nolan) were cultured in regular growth medium consisting of high-glucose DMEM (Invitrogen) with GlutaMAX containing 10% FBS, 0.1 mM nonessential amino acids, 100 IU/ml penicillin, 100 μg/ml streptomycin, 25 mM HEPES (pH 7.2), 25 and 1 mM sodium pyruvate. Retroviral particles to infect murine cells were generated by transfecting Phoenix cells with the retroviral luciferase expression plasmid pMSCV-neo-Luc2 (3) using Lipofectamine 2000 in serum-free Opti-MEM media (both Invitrogen). An equal volume of regular growth media supplemented with additional 10% FBS was added 2-3 hours after transfection. After 16 h, media was replaced with regular growth medium containing 10 mM 30 sodium butyrate. After 8 h of incubation, the medium was again changed to regular growth medium. Viral supernatants were collected 24 h later, filtered (0.45-μm), and then centrifuged at 2,000g for 90 min on 50 μg/ml RetroNectin-coated non-tissue-culture-treated six-well plates.
Protein concentration in the supernatant was determined with a Bradford assay (Bio-Rad). Periodate-aminooxy ligation. Sialic acids were labeled via mild periodate oxidation and 20 detected as previously described (6). In brief, 4x10 5 cells were plated in each well of a 96 well plate. Cells were washed three times with PBS, and then resuspended in 1 mM NaIO4 in PBS at pH 7.4 for 30 min on ice. Cells were washed once and each well was resuspended in 100 μL of FACS buffer (0.5% bovine serum albumin (BSA) in phosphate buffered saline (PBS)). 50 μL of 5 mM glycerol was added to each well, mixed by pipetting, and incubated for 10 minutes on ice. 25 Cells were then pelleted and washed twice with 5% FBS in PBS at pH 6.7. Aldehydes were then labeled by resuspending each well of cells in 200 μL of a solution comprising 0.1 mM aminooxy biotin (Thermo Fisher), 10 mM aniline, and 5% FBS in PBS at pH 6.7, and incubating for 90 min on ice. Cells were washed twice with 5% FBS in PBS at pH 6.7, and resuspended in FACS buffer Genome-wide gene expression analysis. RNA was isolated using an RNeasy kit (Qiagen) and assessed for degradation using a 2100 Bioanalyzer (Agilent). cDNA was generated using the 15 SMART-seq v4 kit (Takara) with oligo(dT) priming, and libraries were prepared with the Nextera XT Library Prep Kit (Illumina). Libraries were pooled and sequenced on a HiSeq-4000 (Illumina).
For the MYC inactivation time course experiments, polyA-selected libraries were prepared by the BGI Group using the BGISeq-500 library construction protocol and were sequenced as pairedend reads of 100 base pairs on a BGISeq-500 platform. Reads were demultiplexed with 20 bcl2fastq, mapped to the mm10 reference genome with STAR(8) using the GENCODE vM16 comprehensive annotations, and then counted using featureCounts from Rsubread(9). Differential gene expression analysis was performed using DESeq2(10) with the FDR set to 0.01. P values were adjusted for multiple hypothesis testing. Gene annotations were accessed with biomaRt(11), and GO term analysis was conducted with topGO(12) after setting a log2 fold-change expression 25 threshold of 1. Mouse glycogene annotations were pulled by referencing the GlycoGene Database(13) through GlyCosmos (https://glycosmos.org/). Glycogene expression was also examined in murine Burkitt's lymphoma (GSE51008) (14) and T-ALL (GSE106078) (15). with 0.2% FA) were driven and controlled by a Dionex Ultimate 3000 RPLC nano system (Thermo Fisher Scientific). An integrated loading pump was used to load peptides onto a trap column (Acclaim PepMap 100 C18, 5 µm particles, 20 mm length, Thermo Fisher Scientific) at 5 µL/minute, which was put in line with the analytical column 5 minutes into the gradient. The gradient was held at 0% B for the first 6 minutes of the analysis, followed by an increase from 0% 35 to 5% B from 6 to 7 minutes, an increase from 5 to 25% B from 7 to 66 minutes, an increase from 25% to 90% from 66 to 70 minutes, isocratic flow at 90% B from 70 to 75 minutes, and re-equilibration at 0% B for 15 minutes for a total analysis time of 90 minutes. Eluted peptides were analyzed on an Orbitrap Fusion Tribrid MS system (Thermo Fisher Scientific). Precursors were ionized using an EASY-Spray ionization source (Thermo Fisher Scientific) held at +2.2 kV relative to ground, the column was held at 40 °C, and the inlet capillary temperature was held at 275 °C. ppm, and 20 ppm were used for first search MS1 tolerance, main search MS1 tolerance, and MS2 product ion tolerance, respectively. Oxidized methionine and deamidated asparagine were set as variable modifications, and carbamidomethylation of cysteine was set as a fixed modification. Cleavage specificity was set to Trypsin/P with 2 missed cleavages allowed. Peptide 20 spectral matches (PSMs) were made against a mouse protein database downloaded from Uniprot. Peptides were filtered to a 1% false discovery rate (FDR) using a target-decoy approach (18), and a 1% protein FDR was applied. Proteins were quantified and normalized using MaxLFQ, and the match between runs feature was enabled. Label free intensity values were log2 transformed and plotted using Perseus version 1.6.2.2 (19). Identified proteins were filtered to plot 25 those that are annotated as localized on the cell surface or secreted, but not cytoplasmic (via UniProt annotations). Significance cutoffs for volcano plots were determined using student's t test with a false discovery rate of 0.0001 and minimum enrichment (S0) of 5.  (27). K-means clustering was used to stratify patients into two groups based on MYC and ST6GALNAC4 expression. Welch's t test was used to compare the proportion of different immune subsets in MYC high / ST6GALNAC4 high tumors (n=1780) versus 5 MYC low / ST6GALNAC4 low tumors (n=9323). Lastly, the correlation between MYC and ST6GALNAC4 expression was evaluated using a Spearman correlation test.
The reaction was stirred for 6 hours at room temperature. The solvent was then evaporated in vacuum and the residue was purified on a Biogel P2 column (Biorad, 1504118) to afford product as a white foam. Spectral data were in accordance with the literature (32). For use in 5 chemoenzymatic reactions, products were dissolved in water to 180 mM, aliquoted, and frozen. murine T-ALL line over 12 hours. Fluorescence integrated intensities were normalized to empty vector transfected WT target cells. (d) Side-by-side comparison of the plots in main Fig. 2g and h      for 48 hours at the indicated concentration (n=3 per group, two-tailed Student's t test). Data presented as mean ± s.d. (c) Representative plot of Siglec-7 ligands on PEER cells after treatment with DMSO or 100 μM 10058-F4 for 48 hours. 10 FIGURE S19: Pan-cancer survival covariance Survival Z-scores for all cancers in PRECOG(24) were extracted after stratifying patients by median expression of the indicated gene. A covariance matrix was then computed using the twenty sialyltransferases and MYC. A high positive covariance (red) indicates that two genes 5 have similar effects on patient survival.  Tables   Table S1. (separate file) Processed RNA-seq data comparing MYC dose in T-ALL. 5 Table S2. (separate file) GO term analysis of MYC on and off T-ALL.  Shotgun proteomics of Siglec-7 immunoprecipitation of T-ALL.