s1

ONLINE SUPPLEMENTAL MATERIAL

Supplemental Methods include information regarding isolation of murine immune cells, flow

cytometry analysis and cell sorting, sample preparation of gene expression analysis, multi-photon

in vivo imaging, immunofluorescence staining of epidermal sheets, microarray and RNAseq

analysis, data set pre-processing and gene subtraction, sample relationship visualization,

differential gene expression, GSEA, ssGSEA, WGCNA, prediction of upstream master regulators,

overrepresentation analysis, and module-based differential expression analysis. Supplemental

Table 1, included in a separate Excel file, lists for the 19 highly preserved F→M WGCNA gene

modules, and the 2 skin-specific B6→129 WGCNA modules: the over-represented Gene

Ontology categories (Supplemental Table 1A), the putative driver genes identified on the basis of

high intra-modular connectivity (Supplemental Table 1B), and the transcription factors predicted

to act as upstream regulators (Supplemental Table 1C). Supplemental Table 2, included in a

separate Excel file, lists the clinical details for the 5 HSCT patients included in this study.

Supplemental Table 3, included in a separate Excel file, contains the results from the pre-

processing performed to test for LC contaminants in LC-replete versus LC-depleted epidermal

samples (Supplemental Table 3A), and lists the intestine- and skin-specific genes subtracted

from the analysis (Supplemental Table 3B). Supplemental Figure 1 shows the visualization of

sample relationship without gene subtraction. Supplemental Figure 2 shows the overlap between

TE gene expression profiles in the F→M and B6→129 experimental models, and the similarities

between F→M and B6→129 based WGCNA modules. Supplemental Figure 3 depicts the

analytical pipeline followed, based upon correlation network analysis and downstream validation.

Supplemental Figure 4 shows the network representation of the B6→129 skin-specific modules

MD and ME, and their overlap with the F→M skin-specific module M28. Supplemental Figure 5

shows the expression of JAG1 and DLL4 Notch ligands by LC compared to non-LC populations

in the epidermis of mice developing GVHD, and the effect of in vitro Notch blockade on the

capacity of male LC to stimulate IFN-g generation by activated MataHari T cells. Supplemental

Figure 6 shows the kinetics of LC host-to-donor turn over and the effect host LC depletion upon

TE accumulation in the epidermis in independent murine BMT models of GVHD.

s2

Supplemental Methods

Isolation of murine immune cells

a) Lymph nodes and spleens – To prepare cell suspensions from spleens and lymph nodes, the

freshly removed organs were mashed and passed through a 40 µm cell strainer; red blood cells

were removed by isotonic lysis with ammonium chloride (ACK Lysing Buffer; Lonza, UK). Cells

were re-suspended in FACS buffer (PBS, 2% FCS, 2 mM EDTA; Lonza UK) for counting and

immunolabelling.

b) Bone marrow – To isolate bone marrow cells, both epiphyses of the long bones of the hind

limbs were cut and the bone marrow was flushed out with FACS buffer. The cell suspension was

filtered through a 40 µm cell strainer and red blood cells were removed by isotonic lysis with

ammonium chloride. Cells were re-suspended in FACS buffer for counting and immunolabelling.

c) Blood – Erythrocytes were removed from whole blood samples by hypotonic lysis with distilled

water. Cells were re-suspended in FACS buffer for counting and immunolabelling.

d) Small intestine – The entire small intestine was flushed and rinsed with 40 ml of ice cold

harvest medium (PBS, 2% FCS, 1% penicillin-streptomycin; Lonza, UK), and sectioned into

~0.5 cm pieces. The intestinal pieces were incubated with detaching medium (HBSS, 5% FCS,

1% penicillin-streptomycin, 5 mM EDTA; Lonza, UK) at 37°C with shaking at 150 rpm for 60 min.

The supernatant, containing the IEL, was passed sequentially through a 100 µm and a 40 µm cell

strainer and enriched for lymphocytes by density gradient centrifugation with Ficoll PaqueTM Plus

(GE Healthcare, UK). The intestinal pieces were further incubated in digestion medium (RPMI,

2% FCS; Lonza, UK; 200 U/ml collagenase IV; LifeTechnologies, USA; 200 U/ml DNAse I;

Sigma, UK) at 37°C with shaking at 150 rpm for 60 min, dissociated and passed sequentially

through a 100 µm and a 40 µm cell strainer. The cell suspension, containing the LP cells was

enriched for lymphocytes by density gradient centrifugation with Ficoll PaqueTM Plus. Cells were

re-suspended in FACS buffer for counting and immunolabelling.

e) Skin – Epidermal and dermal immune cells were isolated from the skin in accordance with the

protocol described by Henri et al. (1), with minor modifications. Briefly, the body skin was cut into

~1x1 cm pieces, after having been shaved and the subcutaneous fat removed, and the ears were

s3

split in two parts (ventral and dorsal). The pieces of skin were incubated overnight in dispase

medium (HBSS, 2% FCS; Lonza, UK; 2.5 mg/dl dispase II; Sigma, UK) at 4°C, and the epidermal

and dermal sheets were separated and cut into ~0.5 cm fragments. The epidermal fragments

were vortexed, mashed and passed sequentially through a 100 µm and a 40 µm cell strainer to

disintegrate the remaining tissue and create a cell suspension. The dermal fragments further

incubated in digestion medium at 37°C with shaking at 150 rpm for 60 min, dissociated and

passed sequentially through a 100 µm and a 40 µm cell strainer. Cells were re-suspended in

FACS buffer for counting and immunolabelling.

Flow cytometry

Cells were plated out at up to 1x106 cells per well in a 96 well conical bottom plate and incubated

with 2.4G2 antibody at 4°C for at least 10 min to block Fc receptors. For cell surface

immunolabeling, cells were incubated with the fluorochrome-conjugated antibodies diluted in

100 μl of FACS buffer at 4°C for 20 min in the dark – CD4 (GK1.5, eBioscience, USA), CD8a (53-

6.7, BD Biosciences, Germany), CD11b (M1/70, eBioscience, USA), CD45 (30-F11, BioLegend,

USA), CD45.1 (A20, BD Biosciences, Germany), CD45.2 (104, eBioscience, USA), Thy-1.1

(HIS51, eBioscience, USA), Thy-1.2 (53-2.1, BD Biosciences, Germany), EpCAM (G8.8,

eBioscience, USA), MHC class II I-A/I-E (M5/114.15.2, eBioscience, USA), DLL4 (HMD4-1,

BioLegend, USA), JAG1 (HMJ1-29, BioLegend, USA), LPAM-1 (DATK32, BioLegend, USA),

Vβ8.3 TCR (1B3.3, BD Biosciences, Germany). When intracellular staining was required, after

having performed cell surface immunolabelling, samples were washed twice with FACS buffer,

fixed in 100 µl of fixation solution (BD Cytofix/Cytoperm solution; BD Biosciences, UK) for 15 min

at 4°C in the dark, washed twice with permeabilization solution (BD Perm/Wash™ buffer;

BD Biosciences, UK), and incubated with the fluorochrome-conjugated antibodies diluted in

100 μl of permeabilization solution at 4°C for 30 min in the dark – Active Caspase-3 (C92-605,

BD Biosciences, Germany), CD207 (eBioL31, eBioscience, USA), IFN-γ (XMG1.2, BD

Biosciences, Germany). For detection of cytokine production, cells were treated with brefeldin A

(Sigma, UK) for 2 h at 37°C, prior to immunolabelling. Samples were washed twice with FACS

buffer and re-suspended in 300 µl of FACS buffer for immediate analysis by flow cytometry. For

s4

dead cell exclusion, 2 µl of propidium iodide (LifeTechnologies, USA) was added to the unfixed

samples prior to analysis. For assessment of cell proliferation, cells were stained with

carboxyfluorescein succinimidyl ester (CellTrace™ CFSE Cell Proliferation Kit, Invitrogen, USA),

according to the manufacturer’s instructions. Multicolor flow cytometry data acquisition was done

with BD LSRFortessa and BD LSR II cell analyzers equipped with BD FACSDiva v6.2 software

(BD Biosciences, Germany). Fluorescence activated cell sorting was performed on a

BD FACSAria equipped with BD FACSDiva v5.0.3 software (BD Biosciences, Germany). All

samples were maintained at 4°C for the duration of the sort. Sort purity was accessed for all

samples and only those with purity ≥ 95% were used for RNA extraction. Cells were sorted

directly into Buffer RLT (QIAGEN, USA) with 1% 2-β-mercaptoethanol (Sigma, UK), disrupted

through vortexing at 3200 rpm for 1min, and immediately stored at -80°C until further processing.

Three biological replicates were obtained for every tissue from 3 independent experiments, each

containing a minimum of 4000 cells (pooling where necessary from multiple mice from individual

experiments). Flow cytometry data were analyzed with FlowJo X v10 (LLC, USA).

Sample preparation for gene expression analysis

RNA was extracted using the RNeasy Micro Kit (QIAGEN, USA) following the manufacturer’s

protocol. RNA yield, quality and integrity were evaluated using the RNA 6000 Pico kit on an

Agilent 2100 Bioanalyser (Agilent Technologies, USA). Only samples with a RNA Integrity

Number (RIN) above 8.0 were included in the study. For microarray studies, the Ovation Pico

WTA System V2 kit (NuGEN, USA) was used to prepare amplified cDNA from total RNA.

Spectrophotometric absorbance of the cDNA products at 260, 280 and 320 nm was determined

to assess purity; all samples had an adjusted (A260 – A320):(A280 – A320) ratio > 1.8.

Fragmentation and labeling of the cDNA samples was performed using the Encore Biotin Module

kit (NuGEN, USA), according to kit instructions, and then hybridized onto GeneChip Mouse Gene

2.0 ST arrays (Affymetrix, USA). Hybridisation was performed in a single batch per experimental

system. For RNA-seq studies, RNA samples were amplified using the SMART Seq® v4 Ultra®

Low Input RNA Kit. Paired-end sequencing libraries were prepared from the amplified cDNA

according to the Nextera® XT DNA library prep protocol, and sequenced (38 base-paired reads)

s5

using an Illumina NextSeq 500 (Illumina, USA). Microarray and RNA-seq studies were performed

in collaboration with UCL Genomics.

Multi-photon in vivo imaging

Mice were anesthetized with fentanyl-fluanisone (Hypnorm®; VetaPharma, UK) 0.4 ml/kg and

midazolam 5 mg/kg injected IP. Anesthesia was prolonged by injecting additional doses of

Hypnorm® (0.3 ml/kg) every 40-60 min. Ears were depilated using NairTM Hair Remover (Church

& Dwight, UK). Mice were then placed on a custom-made imaging platform and the ears fixed

with a double-sided tape. PBS was placed on top of the ear and a water reservoir, created with a

metal ring sealed by a coverslip, was glued on top. Mice were then placed in the thermostated

imaging chamber and movies of 22-60 min were acquired. Images were acquired with the Leica

DM6000-CFS multiphoton microscope (Leica Microsystems, UK) enclosed in a dark chamber

heated at 37°C with heated air, and the HC FLUOTAR 25x/0.95 N.A. water immersion objective

(Leica Microsystems, UK). EYFP and DsRed were excited at 950 nm. The emitted fluorescence

was acquired using a hybrid detector (HyD; Leica microsystems, UK). The second harmonic

generation, produced by the collagen present in the dermis, was visualized by setting the first

HyD channel to a 450-500 nm filter window; the EYFP fluorescence by setting the second HyD

channel to 510-565 nm and the DsRed fluorescence by setting the third HyD channel to 575-

660 nm. Raw image data were processed with Fiji (ImageJ): drift was corrected with the Fiji

plugin “Correct 3D drift” using the SHG channel as reference; a median filter (0.3 pixel) was

applied to reduce background noise and autofluorescence was removed using the command

Image Calculator to subtract crossover signal between channels:

corrected channel 1 = channel 1 – channel 2 – channel 3

corrected channel 2 = channel 2 – channel 1 – channel 3

corrected channel 3 = channel 3 – channel 1 – channel 2

Compilation images (z-stack projection and yz orthogonal view) were created with Imaris 8.3

software (Bitplane).

s6

Immunofluorescence staining of epidermal sheets

Epidermal sheets were fixed in ice-cold acetone, blocked in a 0.25% fish gelatin, 10% goat serum

solution and stained for LC with CD207 (eBioL31) (eBioscience, USA). Images were acquired

with the Leica DMI4000b fluorescence microscope (Leica Microsystems, UK), using the Leica

Application Suite Advanced Fluorescence v3.2 software (Leica Microsystems, UK). Raw image

data were processed with Fiji (ImageJ).

Microarray analysis

A total of 36 RNA samples from the B6®129 model and 61 RNA samples from the F®M model

were used for microarray analysis, representing 3 independent biological replicates for each of

the tissues/compartments, treatment conditions and time points. Hybridized arrays were scanned

with a GeneChip 3000 7G scanner (Affymetrix, USA) (1 batch for B6®129 samples; 3 batches

for F®M samples) and the image data processed to generate .cel files. Expression Console

Software, version 1.4.1 (Affymetrix, USA) was used to generate quality control statistics for each

sample; samples with a high mean absolute deviation of the residual for a chip versus all chips in

the data set (≥ data set average + 2 SD), a high mean absolute relative log expression (≥ data set

average + 2 SD) and a low area under the curve (AUC) for a receiver operator curve (ROC)

comparing the intron controls to the exon controls (≤ 0.8) were considered to have poor

hybridisation quality and excluded from the analysis (1 sample in the B6®129 data set; 3

samples in de F®M data set). Raw sample expression signals were background subtracted,

quantile normalized, and the probe level data were summarized using the Robust Multi-array

Average algorithm (2, 3) implemented in the oligo BioConductor R package (4). The ComBat

algorithm from the sva BioConductor R package (5) was employed to adjust for batch effects.

Transcripts identified through multiple probes were collapsed based on maximum expression

values using the CollapseDataset module of GenePattern software (Broad Institute) (6).

Data set pre-processing and gene subtraction

Pre-processing of the data sets was performed to test for LC contamination in the microarray

samples used in experiments shown in Fig. 8. We found no significant difference between LC-

s7

replete and LC-depleted epidermal samples in terms of the levels of expression of LC signature

genes (7) (Supplemental Table 3A). To exclude any effect of tissue-specific transcripts derived

from non-T cell contamination of individual samples, transcripts identified as highly specific for the

intestine and skin in the PaGenBase database [specificity measure (PMS) > 0.9] (8) were

subtracted from the entire dataset prior to analysis (Supplemental Table 3B).

RNAseq analysis

FASTQ Toolkit, version 1.0.0 (9), was used for adapter trimming of the reads. Alignment and

mapping of all libraries were performed using TopHat Alignment, version 1.0.0 and Cufflinks

Assembly & DE, version 2.0.0, selecting Homo sapiens hg38 RefSeq gene annotations. Gene

expression levels were calculated using the Cufflinks Assembly & DE, version 2.0.0, employing a

geometric library normalization method and a fragment bias correction algorithm.

Samples relationship visualization

Multivariate statistical analysis methods implemented in the stats R package, in particular

multidimensional scaling, were applied to perform dimensionality reduction of the datasets and

visualization of the samples relationships. Similarities between groups were evaluated by

hierarchical clustering, using average-linkage upon the Pearson’s correlation-based dissimilarity

matrix; the validity and stability of the clusters was assessed using a non-parametric bootstrap as

implemented in the pvclust R package (10). Additional hierarchical clustering and heat maps

were produced using the matrix visualization and analysis platform GENE-E (Broad Institute,

USA).

Differential gene expression

The limma BioConductor R package was used to perform analyses of gene differential

expression, using an empirical Bayes moderated t-statistic corrected for multiple hypothesis

testing using Benjamini-Hochberg procedure, with a cut-off of false discovery rate (FDR) ≤ 0.05,

and an absolute fold-change cut-off ≥ 2.0.

s8

Gene set enrichment analysis (GSEA)

Gene Set Enrichment Analysis was performed using the GSEA software with the gene sets

derived from the Biological Process Ontology database (C5) defined by the Gene Ontology

Consortium, collected in the Molecular Signatures Database (MSigDB v5.1), and the gene sets

derived from the modules identified by WGCNA in the F→M model dataset.

Single-sample gene set enrichment analysis (ssGSEA)

ssGSEA was performed using the GSVA R package employing the “ssgsea” method (11), as

described by Hänzelmann et al. (12). The Tc1 gene signature was derived from Best et al. data

set (13) as the top 100 over-expressed genes (fold change ≥ 1.5 and FDR ≤ 0.05) in effector OT-I

cells on day 6 post Lm-OVA or VSV-OVA infection in comparison to naïve OT-I cells; the Tc17

gene signature was derived from Gartlan et al. (14) data set as the differentially expressed genes

(fold change ≥ 1.5 and FDR ≤ 0.05) between CD8+YFP+ and CD8+YFPneg T-cells 7 days after

allogeneic transplant; the MDR1+ Th1/Th17 gene signature corresponds to the ex vivo

transcriptional signature of MDR1+ Th17.1 cells isolated from involved Crohn’s disease patient

tissue published by Ramesh et al. (15) (Supplemental Table 4).

Weighted gene co-expression network analysis (WGCNA)

Scale-free network topology analysis of microarray expression data was performed using the

WGCNA R package, as previously described (16, 17). A signed hybrid weighted correlation

network was constructed using a Pearson correlation matrix created from the pairwise

comparison between all pairs of genes, and a soft thresholding power β=8. The topological

overlap was calculated as a measure of network interconnectedness, and genes were grouped

by average linkage hierarchical clustering on the basis of the topological overlap dissimilarity (1-

topological overlap). Module eigengenes were calculated using a dynamic tree-cutting algorithm

and merging threshold function at 0.25. The modules identified were correlated to the sample

traits using a binary vector representation of the tissues of origin and study groups. To validate

the microarray analysis, preservation of the WGCNA modules identified in the F→M model

dataset was tested against the dataset of the B6→129 model, using the R function

s9

“modulePreservation” in the WGCNA R package, as previously described (18). Results were

interpreted according to the following thresholds for Zsummary: if Zsummary≥10, module “strongly

preserved”; if Zsummary ≥2 and <10, module “weak to moderately preserved”; if Zsummary<2, module

“not preserved”. Significance of the Zsummary scores was calculated by gene permutation testing.

Intramodular hub genes, which are genes that have the highest number of connections within a

module, were identified on the basis of having eigengene-based connectivity (kME) > 0.8 and

gene significance (GS) > 0.2. Visualisation of the modules network of gene connections was

accomplished with the Cytoscape v3.5 software (19).

Prediction of upstream master regulators

The Cytoscape plugin iRegulon was used to predict the transcriptional regulatory network

underlying each of the modules of co-expressed genes, as described by Janky et al. (20). Briefly,

for each gene set, the regulatory sequences in 20 Kb around the transcription start site were

analyzed, and motif prediction was performed using an enrichment score threshold of 2.0, a ROC

threshold for AUC calculation of 0.03%, and a rank threshold for visualization of 5000. Candidate

transcription factors were predicted with a maximal FDR for motif similarity of 0.05.

Overrepresentation analysis

The Web-based Gene Set Analysis Toolkit (WebGestalt), a suite of tools for functional

enrichment analysis, was used to identify overrepresented KEGG pathways categories and

translate gene lists into functional profiles. Statistical significance of KEGG pathways enrichment

was calculated based on hypergeometric distribution statistics, adjusting the false discovery rate

using the Benjamini-Hochberg procedure.

Module-based differential expression analysis (modDE)

In order to quantify potentially significant associations at the level of WGCNA-derived gene

modules between tissue types and/or between conditions, we developed a gene based

association test that takes advantage of the magnitude and sign of the differential expression

effect size for each gene. We start by translating each gene’s p value into a chi-squared statistic

s10

Xi with one degree of freedom. In its simplest form, the test is designed to assess whether genes

are consistently over-expressed in a set of cases compared to controls. Therefore, the test

statistic T is the sum over all genes of the Xi, if Xi is positive, or 0, if Xi is negative. The distribution

of this statistic under the null is obtained by assuming that each gene has a 50% chance of being

over-expressed and 50% of being under-expressed. Note that this can be refined by computing

the genome-wide probability of over-expressed genes, in case that the proportion differs from

50%. With either of these assumptions, the probability of observing a number K of genes over-

expressed among the n genes in the module can be computed using a binomial distribution. For

a given K, the statistic T is chi-squared with K degrees of freedom (because it is the sum of K one

degree of freedom chi-square distributions). The probability of observing a test statistic greater

than the observed value T can then be computed using a weighted sum of probabilities, summing

over all possible values of K. The derivation is analogous if all genes are expected to be under-

expressed. If the expectation is a mixture of over and under expressed genes, the T statistic is

then summed over all genes in a manner that reflects that combination.

s11

Supplemental References

1. Henri S, Poulin LF, Tamoutounour S, Ardouin L, Guilliams M, de Bovis B, et al. CD207+

CD103+ dermal dendritic cells cross-present keratinocyte-derived antigens irrespective of

the presence of Langerhans cells. J Exp Med. 2010;207(1):189-206.

2. Irizarry RA, Bolstad BM, Collin F, Cope LM, Hobbs B, and Speed TP. Summaries of

Affymetrix GeneChip probe level data. Nucleic Acids Res. 2003;31(4):e15.

3. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, et al.

Exploration, normalization, and summaries of high density oligonucleotide array probe level

data. Biostatistics. 2003;4(2):249-64.

4. Carvalho BS, and Irizarry RA. A framework for oligonucleotide microarray preprocessing.

Bioinformatics. 2010;26(19):2363-7.

5. Leek JT, Johnson WE, Parker HS, Jaffe AE, and Storey JD. The sva package for removing

batch effects and other unwanted variation in high-throughput experiments. Bioinformatics.

2012;28(6):882-3.

6. Reich M, Liefeld T, Gould J, Lerner J, Tamayo P, and Mesirov JP. GenePattern 2.0. Nat

Genet. 2006;38(5):500-1.

7. Artyomov MN, Munk A, Gorvel L, Korenfeld D, Cella M, Tung T, et al. Modular expression

analysis reveals functional conservation between human Langerhans cells and mouse cross-

priming dendritic cells. J Exp Med. 2015;212(5):743-57.

8. Pan JB, Hu SC, Shi D, Cai MC, Li YB, Zou Q, et al. PaGenBase: a pattern gene database for

the global and dynamic understanding of gene function. PLoS One. 2013;8(12):e80747.

9. Cock PJ, Fields CJ, Goto N, Heuer ML, and Rice PM. The Sanger FASTQ file format for

sequences with quality scores, and the Solexa/Illumina FASTQ variants. Nucleic Acids Res.

2010;38(6):1767-71.

10. Suzuki R, and Shimodaira H. Pvclust: an R package for assessing the uncertainty in

hierarchical clustering. Bioinformatics. 2006;22(12):1540-2.

s12

11. Barbie DA, Tamayo P, Boehm JS, Kim SY, Moody SE, Dunn IF, et al. Systematic RNA

interference reveals that oncogenic KRAS-driven cancers require TBK1. Nature.

2009;462(7269):108-12.

12. Hanzelmann S, Castelo R, and Guinney J. GSVA: gene set variation analysis for microarray

and RNA-seq data. BMC Bioinformatics. 2013;14:7.

13. Best JA, Blair DA, Knell J, Yang E, Mayya V, Doedens A, et al. Transcriptional insights into

the CD8(+) T cell response to infection and memory T cell formation. Nat Immunol.

2013;14(4):404-12.

14. Gartlan KH, Markey KA, Varelias A, Bunting MD, Koyama M, Kuns RD, et al. Tc17 cells are

a proinflammatory, plastic lineage of pathogenic CD8+ T cells that induce GVHD without

antileukemic effects. Blood. 2015;126(13):1609-20.

15. Ramesh R, Kozhaya L, McKevitt K, Djuretic IM, Carlson TJ, Quintero MA, et al. Pro-

inflammatory human Th17 cells selectively express P-glycoprotein and are refractory to

glucocorticoids. J Exp Med. 2014;211(1):89-104.

16. Langfelder P, and Horvath S. WGCNA: an R package for weighted correlation network

analysis. BMC Bioinformatics. 2008;9:559.

17. Langfelder P, and Horvath S. Fast R Functions for Robust Correlations and Hierarchical

Clustering. J Stat Softw. 2012;46(11).

18. Langfelder P, Luo R, Oldham MC, and Horvath S. Is my network module preserved and

reproducible? PLoS Comput Biol. 2011;7(1):e1001057.

19. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a

software environment for integrated models of biomolecular interaction networks. Genome

Res. 2003;13(11):2498-504.

20. Janky R, Verfaillie A, Imrichova H, Van de Sande B, Standaert L, Christiaens V, et al.

iRegulon: from a gene list to a gene regulatory network using large motif and track

collections. PLoS Comput Biol. 2014;10(7):e1003731.

s13

Supplemental Table 1, included in a separate Excel file. Characterization of the 19 highly

preserved F→M WGCNA gene modules, and the 2 skin-specific B6→129 WGCNA modules: (A)

over-represented GO categories, (B) putative driver genes identified on the basis of high intra-

modular connectivity, and (C) transcription factors predicted to act as upstream regulators. AUC,

area under the curve; FDR, false discovery rate; kME, module eigengene-based connectivity;

NES, normalized enrichment score.

Supplemental Table 2, included in a separate Excel file. Clinical details concerning the 5

patients included in the paired-tissue RNAseq analysis. ALL, acute lymphoblastic leukemia; AML,

acute myeloid leukemia; 2ry AML, secondary acute myeloid leukemia; CR1, first complete

remission; CR2, second complete remission; DLI, donor lymphocyte infusion; Haplo,

haploidentical donor; HL, Hodgkin s lymphoma; MMF, mycophenolate mofetil; MUD, matched

unrelated donor; SIB, sibling donor; TBI, total body irradiation.

Supplemental Table 3, included in a separate Excel file. (A) Results from the pre-processing

performed to test for LC contaminants in LC-replete versus LC-depleted epidermal samples. (B)

List of the intestine- and skin-specific genes subtracted from the analysis. DT, diphtheria toxin;

FDR, false discovery rate; SD, standard deviation.

Supplemental Table 4, included in a separate Excel file. Tc1, Tc17 and MDR1+ Th1/Th17

gene signatures.

Supplemental Figure 1. MDS of SLO and TO-derived TE samples in B6→129 BMT model: no prior subtraction of skin- and gut-derived

gene sets. MDS plot showing the proximity of the complete transcriptional profiles of donor-derived CD8+ T cells isolated from different organs.

donor (SLO)

syn-BMT (SLO)

allo-BMT (SLO)

allo-BMT (TO)

Legend:

Coordinate 1

10x103

0-10x103

0

10x103

5x103

-5x103

Coord

inate

2

-20x103

Supplemental Figure 2. Spatial diversity of TE profiles is conserved between two independent experimental BMT models. (A) Average linkage hierarchical clustering of all samples from B6→129 and F→M BMT models and controls, linked to a similarity matrix with color coded Pearson’s correlation coefficients to show clustering according to experimental system (circles: B6→129; squares: F→M) and sample cohort (light grey: donor; dark grey: syn-BMT; black: allo-BMT). (B) F→M BMT module preservation in B6→129 BMT model showing the summary statistics Zsummary as a function of the module size (red line Zsummary = 2; blue line Zsummary = 10). (C) Correlation matrix depicting the association between the 22 WGCNA gene modules defined in B6→129 dataset (MA-MV) and the experimental groups (donor, syn-BMT, allo-BMT), GVHD subgroups (SLO, TO), and GVHD individual organs. Cell color and cell number indicate Pearson’s correlation coefficient and corresponding -log10(p-value), respectively. (D) Bar graphs showing the mean eigengene expression in each of the tissues, for donor a syn-BMT (MT), pan-GVHD TO (MI), organ-selective (MH, MF, MD) and tissue-specific modules (MJ, ME). (E) B6→129 BMT module preservation in F→M BMT model showing the summary statistics Zsummary as a function of the module size (red line Zsummary = 2; blue line Zsummary = 10).Bl, Blood; Der, dermis; Epi, epidermis; IEL, intraepithelial lymphocytes; LN, lymph nodez; LP, lamina propria; SLO, secondary lymphoid organs; Sp, spleen; TO, target organs.

A B

D

donor+

syn-BMTallo-BMT

SLOallo-BMTGVHD TO

T

Eige

ng. e

xpr.

-0.2-0.10.00.10.20.30.4

Sp LN Sp LN Sp Bl LP IEL Der Epi-0.20.00.20.40.6

H

F

J

D

E

-0.4-0.20.00.20.4

-0.2

0.0

0.2

0.4

-0.20.00.20.40.6

-0.4-0.20.00.20.4

-0.6-0.4-0.20.00.2

I

Dono

r

syn-

BMT

allo

-BM

T

allo

-BM

T: S

LO

allo

-BM

T: G

VHD

TO

Lym

ph n

odes

Sple

en

Bloo

d

Gut

- LP

Gut

- IE

L

Skin

- de

rmis

Skin

- ep

ider

mis

-1 10

Pearson’s correlation

ABCDEFGHIJKLMNOPQRSTUV

0.38 1.19 1.501.92 0.26 1.460.75 0.35 0.891.98 1.92 4.400.23 0.35 0.510.60 1.70 2.311.09 1.65 2.913.53 1.38 4.915.99 0.23 3.120.01 0.74 0.600.80 0.33 0.900.15 1.32 0.820.01 1.07 0.870.22 1.32 0.740.16 1.74 1.090.69 4.62 1.910.18 4.58 2.7418.0 0.38 2.544.90 2.74 10.70.80 6.47 7.951.08 2.00 3.353.07 0.51 2.62

2.00 6.062.76 0.451.46 0.260.12 4.041.36 2.440.81 4.461.75 9.370.46 2.040.01 2.410.09 0.371.59 3.680.71 0.020.01 0.700.21 0.320.06 1.030.30 2.460.66 0.830.43 1.010.30 4.690.47 3.150.75 0.973.55 0.26

0.98 0.38 0.69 1.14 0.07 8.78 0.100.85 0.76 1.06 0.05 0.93 5.26 0.270.07 0.21 2.32 1.58 3.41 1.61 0.820.58 0.01 0.44 0.12 0.25 2.20 3.930.75 0.38 0.36 0.47 0.20 0.60 15.80.54 0.06 0.40 1.32 5.20 0.61 0.251.36 0.19 0.41 0.48 2.25 0.12 3.280.43 1.75 0.16 0.42 0.26 0.02 0.140.23 0.07 0.23 0.38 0.41 0.79 1.050.18 0.29 0.20 0.16 0.11 2.99 0.261.65 0.26 0.83 0.32 0.24 0.69 0.590.55 0.21 0.72 0.12 0.41 0.35 1.320.12 0.38 0.21 0.42 8.63 0.58 0.490.31 0.31 0.29 1.80 1.02 0.15 0.610.26 2.36 1.08 0.60 0.63 0.80 0.401.61 0.19 0.59 0.82 0.08 1.00 0.510.31 0.16 0.34 0.34 0.10 0.23 0.120.06 0.39 0.19 0.12 0.08 0.48 0.280.12 0.12 0.64 1.40 0.82 1.03 0.400.05 0.01 0.85 0.71 0.63 0.95 0.330.03 0.68 0.42 0.29 0.22 0.93 0.010.98 0.97 1.30 0.75 0.06 0.12 3.55

C

BMT model B6→129 F→M

donor syn-BMTGroup

Legend:

allo-BMT

SpleenSpleen

Lymph nodesLymph nodes

SpleenSpleen

Lymph nodesLymph nodes

SpleenBloodBlood

SpleenBone marrow

Gut - LPGut - IELGut - IELGut - LP

Skin - DermisSkin - Epidermis

Skin - DermisSkin - Epidermis

-1 10

Pearson’s correlation

0

20

40

60

80

10 20 50 100 200 500 1000 2000Module size

Pres

erva

tion Z s

umm

ary

2823

14

21

2

30

1611

1031

17

6

1

29

128

3 413

20

7

927

2518

5

242615

19

22

0

10

20

30

40

50

O

H

RF

D

A

U

KJ

M

G

E

I

P

Q

S

LB

CN

VT

10 20 50 100 200 500 1000 2000Module size

Pres

erva

tion Z s

umm

ary

E

Supplemental Figure 3. Analytical pipeline. A summary of analytical process following WGCNA is shown. See Results text for details.

Gene dendrogram & Module colors

0.6

0.8

1.0

Heig

ht

Modules

WGCNA

Identify modules ofco-expressed genes

0 5 10 15 20 25

0.2

0.4

0.6

0.8

bisque4 cor=0.72, p=5.7e−72

Connectivity

Gen

e Si

gnifi

canc

e

Connectivity0 20 40 60 80

0.0

0.1

0.2

0.3

0.4

0.5

0.6 green cor=0.067, p=0.056

Gen

e Si

gnifi

canc

e

0 5 10 15

0.0

0.1

0.2

0.3

0.4

darkolivegreen cor=−0.13, p=0.13

Connectivity

Gen

e Si

gnifi

canc

e

0 5 10 15 20 25

0.0

0.2

0.4

0.6

orange cor=0.39, p=2.9e−12

Connectivity

Gen

e Si

gnifi

canc

e

Relate modulesto array information

Find key driversin interesting modules

iRegulon

Detect master regulons

Predict directTF-target interactions

Whole-transcriptomemicroarray analysis

Murine aGVHD model

WebGestalt

250

5

Notch signaling

Osteoclast differentiation

Cell cycle

MAPK signaling

RIG-I-like receptor signaling

Glycerophospholipid metabolism

B cell receptor signaling

TGF-beta signaling

Glycerolipid metabolism

Adherens junction

Apoptosis

Adipocytokine signaling

NOD-like receptor signaling

PPAR signaling

Cytosolic DNA-sensing

Long-term potentiation

Melanogenesis

Toll-like receptor signaling

T cell receptor signaling

Neurotrophin signaling

0 50 100 150 200

0 1 2 3 4

Ratio of enrichment

FDR [-log10(q-value)]

Ratio of enrichment

FDR

Determineenrichment forbiological pathways

Functional perturbation

to test module function

in vivo

Evaluate module preservationin independent data sets

Supplemental Figure 4. Module M28 (defined in the F→M BMT model) overlaps with 2 epidermis-specific modules (MD and ME) independently identified in the B6→129 BMT model. (A) Cytoscape generated visualization of the network connections among the 100 most connected genes in B6→129 MD and ME. Nodes represent the genes (circle area proportional to the intra-modular connectivity, kME) and the color reflects the FDR q-value of its correlation with the module; edges represent the topological overlap between genes (line thickness proportional to adjacency). Driver genes common to F→M M28 are highlighted in bold. (B) Graphs showing the ratio of enrichment (bars) and FDR q-values (line) for KEGG pathways predicted by WebGestalt to regulate MD and ME. Pathways common to F→M M28 are highlighted in bold. (C) Graph showing module association with immunossupressive therapy resistance assessed by determining the over-representation of a gene signature specific for a human MDR1+ Th1/Th17 subset that is resistant to glucocorticoids. Hypergeometric test.FDR, false discovery rate; kME, intra-modular connectivity.

B

Osteoclast differentiation

Wnt siganling pathwayNotch signaling pathway

Cell cycle

MAPK signaling pathway

1500 50 100Ratio of enrichment

60 2 4FDR [-log10(q-value)]

Module D

1500 50 100Ratio of enrichment

Osteoclast differentiation

Toll-like receptor signaling

MAPK signaling

Jak-STAT signalingB cell receptor signaling

TGF-beta signaling

T cell receptor signaling

Cell cycle

Chemokine signalingAdipocytokine signaling

Adherens junction

ErbB signalingNeurotrophin signaling

Notch signaling

Cytosolic DNA-sensing

NOD-like receptor signaling

RIG-I-like receptor signaling

Long-term potentiation

150 5 10FDR [-log10(q-value)]

Ratio of enrichment FDR

Module E

-log1

0(p-

valu

e)

Modules AB

CD

EF

GH

IJ

KL

MN

OP

QR

ST

UV

0

1

2

3C MDR1+ Th17.1 signature

Module E

Rara

Egln3

Ski

Ncf4

Sorl1

Styk1

Dscam

Fhod1

Nckap1

Tiam1

Trpv2

Ifitm3

Nfatc2

Swap70

Efemp2

Pvr

Mapkapk3

Sit1

Soat1

Nek7

Gpr68

Isg20

Rfx5

Ubash3a

Abcc1

Cep170b

Ebpl

Ccr8

Ly6g5b

Adgrg3

Cd101

Il13

Cdh1

Rasgrp4

Tjp1

Qpct

Gdpd5

Sema6d

Pld3

Lmna

Fam20a

Csf1

Lta

Rhbdf2

Dkkl1

Athl1

Itsn1

Cmklr1

Cd44

Pag1

Trpm6

Ptgfrn

Ptafr

Fndc3a

Anxa1

Dnase1l1

Pxn

P2ry14

Fut7

Rbpj

Slc9a9

Itpripl2

Ifitm1

Smpd1

Med10

Plec

Capg

Ldlrad4

Fam129b

Ar

Mgat5

Hrh4

Rnf13

Rab3d

Kif3a

Ahr

Tnf

Cobll1

Myadm

Rxra

Synj2

Adgrg1

Cysltr1

A

Clic1

Cnih1

Iqgap1

Pqlc3

Tab2

Zmiz1

Dstn

Tspan2

Car5b

Cd99l2

Mif4gd

Bscl2

Atp10a

Thy1

Ccr2

Pacsin2

Hopx

Arl15

Prr13

Tjap1

Rell1

Actr1a

Golt1b

Krtcap2

Itgax

Slc2a1

Tpm4

Cd40lg

Ifng

Plp2

Ptprj

Galnt3

Vim

L1cam

Anxa2

Dennd5a

Adam8

Ptpn13

Reep5

Fam188a

Lgals3

Ahnak

Degs1

Crip1

Aim1

Lgals1

Mapkapk2

Atf6

Ctsd

Ulbp1

Klrk1

Il2ra

Id2

Crip2

Capn2

Cd82

Gna15

S100a4

Adipor2

Fam129a

Anxa5

Dgkh

Havcr2

Map2k3

Gata3

Fut8

Klrc1

Ppp1r11

Dok2

Anxa4

S100a10

Ifitm2

Prelid1

Rbms1

Serinc3

Plxdc1

Pcgf2

Raph1

Gcnt1

Diaph1

Emp1

Surf4

Ero1l

Prex1

Lmf2

Carhsp1

C3orf58

Module DCircle area = kME

Color intensity = FDR

0.60.81.0

>0.05<0.01<0.001

Supplemental Figure 5. LC express higher levels of JAG1 and DLL4 than other epidermal cell populations. (A) Notch ligand expression by the main cell populations in the epidermis post-BMT. Top - Representative flow cytometric plots showing JAG1 and DLL4 expression by LC versus CD4+/γδ T cells versus keratinocytes in the absence (BM only) or presence (BMT + T cells) of GVHD. CD8+ T cells infiltrating the epidermis in GVHD were excluded by gating. Bottom - Summary data of the difference in MFI staining between each sample and the respective isotype control (n = 2, graphs show mean ± SD; §, isotype control MFI > sample MFI). No detectable DLL1 and JAG2 expression was found on any epidermal population (data not shown). (B) Effect of in vitro LY411575 or PBS (untreated) exposure upon IFN-γ generation by concanalavin-preactivated MataHari T cells upon interaction with male or female LC. Data derived from 6 independent experiments, graphs show mean ± SD. ****p ≤ 0.0001, two-way ANOVA with Holm-Sidak correction for multiple comparisons.BM, bone marrow; F LC, female Langerhans cells; M LC, male Langerhans cells; MFI, median fluorescence intensity.

T cells +M LC

T cells +F LC

T cells

% IF

N-γ

+

0

10

20

30

40 UntreatedLY411575

****B

IFN

-γ

CD8

T cells

T cells+

Female LC

T cells+

Male LC

Untreated LY4115750.9

0.6

0.5

0.2

21.3 1.3

MFI

Keratinocytes

LC

CD4+

&γδ T cells

coun

t

JAG1

coun

t

DLL4

BM only BM + T cells BM only BM + T cellsASampleIsotype control

BM onlyBM + T cells

JAG1

Kerat.LC CD4+

& γδT cells

DLL4

Kerat.LC CD4+

& γδT cells

0

200

400

600

800

Δ M

FI (s

ampl

e - i

soty

pe c

ontro

l)

§ §§

Supplemental Figure 6. Host-derived LC remain the dominant LC population in the epidermis in the first two weeks post-BMT and their

presence is required for epidermal TE accumulation. (A) Kinetics of replacement of host-derived LC by donor-derived LC after BMT.

Graphs showing the relative proportion (top) and absolute number (bottom) of host- and donor-derived LC over the first 4 weeks post-transplant. Baseline (n = 3), week 1 (n = 7), week 2 (n = 7), week 3 (n = 8), week 4 (n = 8). Lines indicate mean (top graph) and geometric mean (bottom graph). (B) Effect of presence or absence of host LC, upon CD8+ TE accumulation in the epidermis in the B6→129 BMT model (left) and CD4+ TE Marilyn accumulation in the F→M BMT model (right). B6→129, n = 8-9/group; Marilyn F→M, n = 6/group. Bars indicate mean + SD. *p ≤ 0.05, two-tailed Mann-Whitney test. LC, Langerhans cells.

ARecipient LC

Donor LC

0

20

40

60

80

100

% L

ange

rin+ E

pCAM

+ cel

ls

B

0 1 2 3 4

102

103

104

105

106

Weeks

Lang

erin

+ EpC

AM+ c

ells

/ g

0

B6 → 129

PBS DT0

20

40

60

80

% D

onor

CD

4+ T c

ells *

MarilynF → M

0

1

2

3

% D

onor

CD

8+ T c

ells

*

PBS DT