Fig 1.
Overview of tissue samples for this study, mentioning the tissue samples used for MMR and p53 status and expression profiles separately.
ESCC in HL survivors (hESCC) will be compared to sporadic ESCC (sESCC). For each group (HL survivors, AUMC and TCGA), non-neoplastic esophageal squamous tissue analyzed using the same platform (NanoString, Illumina sequencing), will be used as a reference for the analysis of the expression profiles of tumor tissue. This will be treatment naïve NN-tissue of sporadic cases (sNN-tissue) for the NanoString profiles and the Illumina sequencing profiles of AUMC and TCGA. NN-tissue of HL survivors (hNN-tissue), which had not been treated for ESCC with chemo- or radio-therapy, will also be analyzed by NanoString, as a comparator group.
Table 1.
Baseline characteristics RNA expression series of ESCC in Hodgkin lymphoma (HL) survivors (hESCC), sporadic ESCC (sESCC) of Amsterdam University Medical Center (AUMC) and sESCC of TCGA.
Fig 2.
Heatmap of differentially activated pathways when comparing tumor tissue of the three different series versus treatment naïve non-neoplastic tissue from sporadic cancers (sNN-tissue) reference.
Of the Pancancer IO 360 Gene expression Panel, 242 genes were differentially expressed between hESCC and the non-neoplastic squamous control tissues (sNN); 610 between the sESCC AUMC and the sNN and 325 between the sESCC TCGA and sNN tissues. These differentially expressed genes were used as input to perform the IPA analysis. In general, similar pathways are (in)activated in the three different patient series (HL survivors, AUMC, TCGA) comparing neoplastic and sNN-tissue. Deactivated pathways in all three different series are PPAR (peroxisome proliferator-activated receptor) signaling and LXR/RXR activation. Activated pathways are immune-related pathways, HMGB1 signaling and colorectal cancer metastasis signaling and ILK signaling.
Fig 3.
Principal component analysis (PCA) plot showing principal component 1 (x-axis) and 2 (y-axis) scores for the different sample types analysed with NanoString.
Non-neoplastic tissue from HL survivors (hNN-tissue) is distinguishable from non-neoplastic tissue from sporadic ESCC cases (sNN-tissue) and ESCC in HL survivors (hESCC) on this PCA plot. PCA plots from sporadic ESCC are shown in a separate figure, since RNA sequencing was performed, which is not comparable to NanoString profiling (S2 Fig in S1 File).
Fig 4.
Heatmaps showing hierarchical cluster analyses results when applied to NanoString profiles.
Cluster analyses on one ESCC in HL survivors (hESCC) and non-neoplastic tissue from HL survivors (hNN-tissue) (Fig 4A) and 2. hESCC and sNN-tissue (Fig 4B) result in separate tumor and non-neoplastic tissue groups.
Fig 5.
Heatmap of differentially activated pathways when comparing ESCC in HL survivors (hESCC) versus non-neoplastic tissue from HL survivors (hNN-tissue).
The z-scores of the significant activated/deactivated pathways estimated by ingenuity pathway analyses are visualized. The “Role of BRCA1 in DNA damage response” is the only significantly activated pathway in hESCC.
Fig 6.
Heatmap showing the 94 differentially expressed genes with similar expression in non-neoplastic tissue from HL survivors (hNN-tissue) and ESCC from HL survivors (hESCC), when compared to non-neoplastic tissue from sporadic ESCC (sNN-tissue).