Common clonal origin of central and resident memory T cells following skin. immunization

Transcription

1 SUPPLEMENTARY INFORMATION Common clonal origin of central and resident memory T cells following skin immunization Olivier Gaide, Ryan O. Emerson 2, Xiaodong Jiang, Nicholas Gulati 3, Suzanne Nizza, Cindy Desmarais 2, Harlan Roi 4, James G. Krueger 3, Rachael A. Clark and Thomas S. Kupper Supplementary Figure. Uniased tracking of T cell clones and their progeny during three immunization schemes, using high-throughput TCRβ sequencing. a, Outline of experimental plan for sampling skin and lymph node tissue for DNA extraction and TCR sequencing efore and after immunization with three distinct antigenic challenges: OVA+CT, or MVA (n=3 mice for each condition, all experiments performed 2-3 times, representative data is shown)., Ear swelling day after the second OVA+CT immunization, immunization or 6 days after MVA infection, respectively. n=3 mice for each group. All experiments performed 2-3 times, representative data shown. c, Frequency (y axis) of respective V(x-axis)J(z-axis) sugroups in the spleen of 3 wild-type C57Bl/6 mice and OT-I mice, highlighting the presence of a single TCR clone in the OT-I mouse ( n=4). d, Relative and asolute numers of the OT-I TCR sequence measured in the LN of mice day after adoptive trafer of E 4, E 5 or E 6 naive OT-I cells, showing dose-dependent aundance of the clone (n=3 mice). As a control, normal C57Bl/6 littermates that were not injected (No T) or the spleen of an OT-I mouse were analyzed (n=9 mice). e, Dot plots (TCR sequencing, each dot represents a unique CDR3 sequence) showing the frequency of TCRβ sequences in the skin and LN of a C57Bl/6 mouse injected with 5 OT-I cells efore (left panel, ILN vertical axis, tail skin horizontal axis) and 6 weeks after (right panel, draining Nature Medicine: doi:.38/nm.386

2 LN y axis, treated skin x axis) immunization with OVA+CT (see also figure ). The OT-I clone CASSRANYEQYF, is present only in the LN efore immunization, as naive OT- cells were injected one day prior, and is detected in oth the skin and LN only after immunization. (n=3 mice). f, Dot plot of the asolute numer of T cells (rather than percentage) oserved in 4ng of gdna from skin and LN, respectively, after OVA+CT immunization. Gray lines link individual clones (dots) etween skin and LN, demotrating that T cell clones are present in comparale asolute numers in oth compartments. The iert shows the numer of OT-I cell clones in the spleen of an OT-I mouse, demotrating only a single sequence (n=3 mice). i, Anatomic tracking of the TCRβ sequences identified in (f), showing a common pattern of asolute T cell numers in the skin and LN, oth draining and distant, occurring after OVA+CT immunization. The OT-I TCR is the sole clone detected in LN prior to immunization, a result of naïve OT-I T cell injection prior to antigen exposure. Numers in skin vs LN are similar, well within an order of magnitude, per 4ng gdna. Supplementary Figure 2. Peripheral TCR presence is not due to tissue contamination y lood T cells. a, Tracking of the OT-I TCR sequence (DNA) in different mice tissues, efore and after OVA+CT immunization (immunization as shown in figure a). The OT-I clone, 24 hour after injection in the lood stream, is not detected in the skin (pre-imm), suggesting that lood contamination of skin undetectale. After immunization, the OT-I clone can e found at the expected sites, i.e. local and distant skin sites and LN, at low levels in lung, ut not in the gut or vagina. If the presence of the clone in distant skin and LN was due to contamination of the tissues y lood, it should also e oserved in the gut or vagina. Detection of the clone in the lung is expected, as skin immunization is known to induce a small population of T cells with lung-migrating properties 2 ). ( n=3 mice)., Blood vs. tissue distriution of the most aundant T cell clones in lood and skin, respectively, in mice repeatedly exposed to (Fig 2). Of the most prevalent T cell clones in the lood, 9 are not detected in the skin or LN (right panel). Nature Medicine: doi:.38/nm.386

3 Conversely, the most prevalent clones in the skin, which are oserved at distant skin and LN sites, are either asent or underrepresented in lood (left panel). Together, these findings demotrate that lood contamination of skin or LN was undetectale, and certainly too low to significantly alter the TCR profiling of mouse tissues. (n=3 mice). Supplementary Figure 3. Frequency of CD4 and CD8 T cells in local and distant skin sites and LN, after multiple immunization. a, Seitization and challenge scheme. Experiments performed 3 times, representative data shown., Dot plots (FACS) of CD4+ and CD8+ in the tail (not exposed), the right ear (exposed once during challenge) and left ear (exposed 4 times during immunization, oost and challenge), with a non-immunized ear skin as control ( n=3 mice). c, Dot plots of CD4+ and CD8+ cells in the LN of immunized and non-immunized mice (n=3 mice). d, Ratio of CD4+ and CD8+ cells, in relation to gamma-delta + cells, in the skin of naïve and immunized skin, either non exposed, challenged once, or seitized and challenged (n=3 mice; values are given as mean ± SD). e, Dot plot of skin-resident TCRβ+ T cells of OVA+CT immunized mice, showing a CD8+ and CD8- (ie CD4+) population (memory phase) (n=3). Supplementary Figure 4. Phenotype of CD8 T cells present in LN (T CM ) and skin (T RM ) days after immunization (memory phase). a, Dot plots (FACS) showing the expression pattern of the indicated surface markers. The expression pattern is the same as that oserved after MVA or VV vaccination, i.e. CD44high, CD62neg, CD69+, CD3 high CD22 neg, E and P-selectin high for T RM (showing antigen experience and skin homing characteristics), and CD62pos, CD69-, E and P-selectin lower for T CM. Representative dot plots of of 9 mice in 3 independent experiments. Nature Medicine: doi:.38/nm.386

4 Supplementary Figure 5. Activated naïve T cell clones give rise to oth skin-resident T RM and recirculating T CM. a, Experimental plan: at day -7, skin and LN iopsies were taken from two mice that would e later e joined y paraiotic surgery. One mouse was seitized on day,, and 7, while the other was not. At day, these mice were surgically joined (paraiosis) and kept as pairs for 4 weeks. At day 63, mice were separated and skin (tail and ear) and LN (inguinal and cervical) were harvested for HTS (n=3 pairs of mice). Experiments were performed 2-4 times., Ear swelling day after challenge in a different cohort of mice at day confirming induction of T RM memory to at this time point (n=6 mice for each condition). c, Dot plots of the frequency (numer of a given sequence divided y the total numer of sequences oserved in the sample) of TCRβ sequences found in the inguinal LN and the skin efore and after paraiosis, with the seitized mouse on the y axis, naïve mouse on the x axis. Prior to paraiosis, syngeneic mice share very few common TCR sequences (<.5% average, 95%CI.3-%). After paraiosis, the numer of common sequences increases dramatically (3% average, 95%CI 2-2%). This increase is roughly equal the numer of sequences common to 2 different LN of the same mouse (~3%; the iert shows the typical overlap of two different LN of a single seitized mouse), demotrating free and complete exchange of T CM etween the circulatory systems of the two mice. Prior to paraiosis, there are no shared clones in skin (lower left); only three clones are evident after paraiosis (lower right), suggesting little if any recirculation of TRM (n=3 pairs of mice). d, Numer of shared T cell clones per 4ng tissue DNA in the LN and skin, efore and after paraiosis, in three independent pairs. Experiments performed twice, representative data shown. The very small rise in the numer of common sequences in skin after paraiosis may represent recirculation of a suset of skin T RM ; alternatively, they could e T EFF or T RM generated during the common (paraiotic) life of the two animals, which would e expected to seed in the skin of oth animals. n=3 pairs of mice. Values are given as mean ± SD. e, Numer of shared T cell clones per 4ng tissue DNA in inguinal LN of paraiotic pairs (mice H and H7, mice H3 and H9, mice H5 and H). f, Dot plot of VβCDR3 sequences found in Nature Medicine: doi:.38/nm.386

5 the exposed skin (ear, y axis) and non-exposed skin(tail, right axis) of a seitized mouse. In green, sequences found only in the ear; in red, sequences only found in the tail. In lue, common sequences, including a TCR defined y the CDR3 peptide CASSDRGGYNSLYF, which is highly expanded in the skin after immunization (n=3 pairs of mice). g, CASSDRGGYNSLYF sequence counts in tissues of oth a seitized mouse and its naïve paraiont. After paraiosis, the OT- clone is found at high frequency solely in the LN of oth naïve and seitized paraionts, identifying it as a T CM, while its presence in the skin of only the seitized paraiont identifies it as a T RM. (n=3 pairs of mice). Supplementary Figure 6. TCR sequence overlap in lymph nodes and skin at different ages (naïve and antigen experienced). Experiments performed twice, representative data shown. a, Left panel: Quantification of the T cell overlap (% unique T cell clone) in the cervical vs. inguinal LN in individual naïve mice (6 week old) and DFNB-seitized mice (4 weeks post, week old). Right panel: T cell overlap in the left vs. right ear in the same mice. Left y-axis, proportion of unique T cell clones oserved in oth samples. Right panel y-axis, proportion of total T cells oserved in oth samples accounting for the frequency of each unique clonotype. Naïve and seitized mice have the same overlap in the numer of unique CDR3 sequences/t cells etween lymph nodes, roughly %. Seitized mice have slightly more overlap etween the two lymph nodes when weighted y total TCR aundance (5% vs. 7%; p =.3). Right panel, TCR overlap at the level of unique sequences is much higher in ears than in LN sample, and higher in seitized than naïve mice (5% vs. %; p =.4). Seitized mice also have sustantially more overlap etween the two ears when weighting y TCR aundance (% vs. 55%; p =.) (n=3 mice for each group)., Tracking of expanded ear-resident T cell clones in other tissues, naïve vs. DFNB-seitized mice. The numer of T cell clones expanded ( 5 T cells in 4ng of total genomic DNA) in oth the left and right ears of naïve and DFNB-seitized mice thresholds that correspond to the selection criteria used to define such expanded clones in Figure. We also calculated Nature Medicine: doi:.38/nm.386

6 the mean clonal size (in T cells per 4ng gdna) of expanded clones, and the proportion of such clones also present in tail, and in cervical and inguinal LN. In contrast to the low overlap oserved among T cells etween different tissues in general (% of unique T cells, as shown aove), T cell clones which are selected ased on high frequency in two skin samples are likely to e found recurrently throughout the animal, and are much more aundant after seitization than efore (n=3 mice for each group). c, Distriution of T cell clonal aundances in 4ng of genomic DNA is shown for naïve and DFNBseitized mice, for skin samples (average of left ear, right ear and tail skin samples). d, Distriution of T cell clonal aundances in 4ng of genomic DNA is shown for naïve and DFNB-seitized mice, for LN samples (average of cervical and inguinal LN samples). n=3 mice in each group. e, Distriution of T cell clonal aundances in 4ng of genomic DNA is shown for DFNB-seitized mice, for LN samples (average of cervical and inguinal LN samples), coidering only T cell clones oserved at high aundance (>5 cells) in at least two skin samples (with an average of 8 T cell clones matching these criteria per mouse, matching the 24 clones reported in panel ). DFNB seitization leads to a large increase in the numer of highly-expanded T cell clones detected in skin samples, and a smaller increase in the numer of expanded T cell clones oserved in LN samples. T cell clones at high aundance in two skin samples are much more likely to e present and highly aundant in LN samples. n=3 mice in each group. All values are given as mean ± SD. Supplementary Figure 7. The rapid memory respoe to is independent of T CM migration, while the slow memory respoe is T CM migration dependent. a, Seitization and treatment scheme for each of the panels elow. Experiments performed 2-3 times, representative data shown., Ratio of the peripheral lood CD3 counts in FTY-7 vs untreated mice, showing the kinetics of the lymphoid sequestration induced y FTY-7. By the 3 rd day of treatment, lood levels of CD3 cells in FTY-7 treated mice were negligile, this persisted through the next four days of treatment, and egan to Nature Medicine: doi:.38/nm.386

7 increase only upon cessation of treatment (n= 3 mice per group). c, FACS analysis of the effect of FTY- 7 on CD4 and CD8 in the lood at day 3 after injection ( n=6). d, Experimental scheme in 4a/II. Left ear swelling after challenge with mediated y adoptively traferred T CM cells is locked y pretreatment with FTY-7, showing a dependence on migration from the lymph nodes ( n= 6 mice per group). e, Right ear swelling after challenge with mediated y adoptively traferred T CM cells is similarly locked y the drug FTY-7 (n=6 mice per group). f, Experimental scheme in 4a/III. Left ear swelling after challenge with is not locked y FTY-7, as it does not require T CM migration (n=6 mice per group). g, Left ear thickness at indicated time points during seitization and challenge with, in FTY-7 and vehicle treated mice, as show in (f) (n=6 mice per group). h, Experimental scheme 4a/IV. Left and right ear swelling after challenge with are not affected y FTY-7 (n=5 mice per group). i, The time to maximal left and right ear swelling after challenge with is not affected y FTY-7 (n=5 mice per group). j, Experimental scheme 4a/V. Mice seitized only once ut not oosted respond slowly to a late challenge, suggesting the asence of significant T RM skin seeding, and a typical T CM respoe ( n=6 mice per group). k, Time to maximal ear swelling after challenge is prolonged in mice seitized ut not oosted ( n=6 mice per group). l, Experimental scheme 4a/VI. Ear swelling after challenge with in mice seitized ut not oosted is FTY-7 seitive, a finding compatile with a role for T CM (n=6 mice per group). m, Maximal ear swelling after late challenge with in mice seitized ut not oosted is FTY-7 seitive ( n=6 mice per group). n, Experimental scheme 4a/VII. Kinetic of ear swelling after late challenge is affected y exposure to FTY-7 during seitization, suggesting that T cell colonization of the skin during seitization is essential to the rapid memory respoe ( n=3 mice per group). o, Time to maximal ear swelling after challenge is prolonged in mice exposed to FTY-7 during seitization (n=3 mice per group). All values are given as mean ± SD. Nature Medicine: doi:.38/nm.386

8 Supplementary Figure 8. T cell respoe kinetics of T RM and T CM are different yet oth are long lasting. a, Seitization, challenge and adoptive trafer scheme. Experiments were performed twice, representative data is shown., Mice exposed to low seitizing doses of at day, 2 and 7 develop a rapid (24h) ear swelling when challenged with at day 3 (n=5 mice per group). c, d, This rapid respoe is also oserved when the mice are challenged at day 6 (c) or day 9 (d) respectively, demotrating that this is a ona fide memory respoe. (c, n=6 per group; d, n=3 per group). e, Mice receiving LN-derived T cells y i.v. adoptive trafer from seitized and oosted mice ( days previously) also develop a significant ear swelling, yet peak swelling is attained much later (5-7 days, n=3 for naïve, 6 for se). f, Time to maximal swelling takes day in the seitized mice (, c and d), ut more than 5 days in mice sujected to adoptive trafer (e, n=6 mice per group). g, Adoptive trafer of freshly seitized mice (D7 post seitization) allows for a fast respoe, likely due to T EFF still present in the draining LN. At a later stage (D), the LN contain only T CM, and their adoptive trafer leads to a slow respoe, as oserved in paraiotic mice (Fig 4; n=6 mice per group). h, Time to maximal swelling of mice shown in (g). n= 6 mice per group. All values are given as mean ± SD. Supplementary Figure 9. Repeated antigenic skin exposure increases the inteity of the T RM respoe. a, Seitization and challenge scheme. Experiments performed 2-3 times, representative data shown., Ear thickness after late challenge with on the left and right ear, in mice seitized and oosted on the left ear, is highest at the skin site previously exposed to the seitizer (left ear), although oth are equally fast (n=5 mice). c, Ear thickness during seitization (on the left ear), oost (on the left ear) and challenges at day (on oth ears), 6 (oth ears) and 9 (oth ears) shows progressive increase of the respoe inteity and kinetics matching the numer of exposure, until the maximal respoe is otained (n=3 mice). d, Ear thickness measures at precise time points efore seitization, Nature Medicine: doi:.38/nm.386

9 during seitization and after each oost, as descried in c. (n=3 mice). All values are given as mean ± SD. Supplementary Figure. TCR clone profiling during DPCP mediated contact dermatitis. a, skin samples from two different sujects share asolutely no CDR3 sequences (left panel). Within the same patient, the middle panel shows placeo treated skin (horizontal axis, red dots), day 3 post DPCP skin (vertical axis, green dots), and shared clones (lue dots). The right panel shows similar data, with placeo treated skin again on the horizontal axis (red), 4 month post DPCP skin on the vertical axis (green), and shared clones (lue). As this technique will yield false negatives ut not false positives, the true numer of shared TCR sequences is likely to higher than what is shown ( n=3 sujects)., Dot plots of the of TCRβ CDR3 sequences shared (or not) in placeo exposed or DPCP exposed skin. Several clones are present at high levels in all samples (lue ox, a-c), irrespective of challenge with DPCP (placeo skin (non exposed), challenged skin at day 3, 4 and month 4). Because they are present at the four month time point in the skin, at high aundance, these T cells are likely to e T RM specific for an environmental antigen encountered efore the trial (not DPCP), as such aundant T RM have een descried efore in unpertured normal human skin 6. Several clones (green ox) are undetectale in placeo skin (non exposed) or challenged skin at day 3, ut present at high frequency at day 4 and month 4 post challenge. These cells likely correspond to newly activated T cells recruited to the skin in respoe to DPCP. (n=3). c, Graphic representation of unrelated TRM clones (a,,c, lue lines) and newly recruited clones (-4, green lines); TCR counts are shown on the y axis, log scale. Clones -4 at 4 months also represent new TRM. (n=3). Nature Medicine: doi:.38/nm.386

10 a Day -8 Day -7 Day = immuniz. Day OVA+CT MVA OVA MVA +OTI iv OVA+CT MVA ss Ear Thickness (xum) 4 p=. OVA 5 4 p=.5 PBS MVA Day Day 8 Day Day 8 Tail skin, left inguinal LN Tail skin, ear skin, cervical LN, right inguinal LN c C57Bl/6 C57Bl/6 2 d No T No T TRBV- TRBV2- TRBV3- TRBV4- TRBV5- TRBV6- TRBV8- TRBV- TRBV2- TRBV2-2 TRBV2-3 TRBV3- TRBV3-2 TRBV3-3 TRBV4- TRBV5- TRBV6- TRBV7- TRBV9- TRBV- TRBV2- TRBV22- TRBV23- TRBV24- TRBV26- TRBJ- TRBJ-2 TRBJ-3 TRBJ-4 TRBJ-5 TRBJ-6 TRBJ-7 TRBJ2- TRBJ2-2 TRBJ2-3 TRBJ2-4 TRBJ2-5 TRBJ2-7 TRBJ-5 TRBJ-52 TRBJ-53 C57Bl/6 3 TRBV- TRBV2- TRBV3- TRBV4- TRBV5- TRBV6- TRBV7- TRBV8- TRBV- TRBV2- TRBV2-2 TRBV2-3 TRBV3- TRBV3-2 TRBV3-3 TRBV4- TRBV5- TRBV6- TRBV7- TRBV9- TRBV- TRBV2- TRBV22- TRBV23- TRBV24- TRBJ- TRBJ-2 TRBJ-3 TRBJ-4 TRBJ-5 TRBJ-6 TRBJ-7 TRBJ2- TRBJ2-2 TRBJ2-3 TRBJ2-4 TRBJ2-5 TRBJ2-7 TRBJ-5 TRBJ-52 TRBJ-53 TRBV- TRBV2- TRBV3- TRBV4- TRBV5- TRBV6- TRBV7- TRBV8- TRBV- TRBV2- TRBV2-2 TRBV3- TRBV3-2 TRBV3-3 TRBV4- TRBV5- TRBV6- TRBV7- TRBV9- TRBV- TRBV2- OT-I TRBV22- TRBV23- TRBV24- TRBV26- TRBJ- TRBJ-2 TRBJ-3 TRBJ-4 TRBJ-5 TRBJ-6 TRBJ-7 TRBJ2- TRBJ2-2 TRBJ2-3 TRBJ2-4 TRBJ2-5 TRBJ2-7 TRBJ-5 TRBJ-52 TRBJ-53 Adoptively trafered cells (numer) No T No T No T No T No T No T No T E4 E5 E4 E5 E6 OTI Asolute OT-I TCR counts p<. Source (OT-I spl.) : E6 OTI E-4... Relative OT-I TCR (%) TRBV- TRBV2- TRBV3- TRBV4- TRBV5- TRBV2- TRBV2-2 TRBV3- TRBV3-2 TRBV3-3 TRBV4- TRBV5- TRBV6- TRBV7- TRBV9- TRBV- TRBV2- TRBV22- TRBV23- TRBV24- TRBV26- TRBV29- TRBV- TRBV3- TRBJ- TRBJ-2 TRBJ-3 TRBJ-4 TRBJ-5 TRBJ-6 TRBJ-7 TRBJ2- TRBJ2-2 TRBJ2-3 TRBJ2-4 TRBJ2-5 TRBJ2-7 TRBJ-5 TRBJ-52 TRBJ-53 e efore immunization TCR percentage (specific/total in sample) after immunization Inguinal LN, pre-imm. (Day-7) CASSRANYEQYF Tail skin, pre-immune (Day-7) Cerv. LN post OVA to ear (D) CASSRANYEQYF Ear skin, 6 wks post OVA+CT to ear (lue= common sequences) f CASSRANYEQYF g Ear 6 weeks after OVA to ear CASSRANYEQYF Tail 6 weeks after OVA+CT to ear Draining LN 6 weeks after OVA CASSRANYEQYF OT-I spleen CASSRANYEQYF Distant LN 6 weeks after OVA+CT to ear Estimated no. of cells in 4 ng of gdna LNT Tail C Ear E Drain. D Tail LN T Dist. LN C Pre- Postimmunization CASSPDKYYAEQFF CASGGGGSYEQYF CASSTNSDYTF CASSDTRGQETLYF CASSLERLGLGDQDTQYF CASSRANYEQYF Supplementary Fig.. Uniased tracking of T cell clones and their progeny during 3 immunization schemes, using high-throughput TCR! sequencing. Nature Medicine: doi:.38/nm.386

11 a Relative OT-I TCR (%)... LN Skin Local Skin Distant Skin Draining LN Distant LN Lung Gut Vagina Pre-imm. days p. immunization (ova/ct) most prevalent TCRs in lood (24h) most prevalent TCR in the ear (24h) Numer of T cells per 4ng of tissue Blood - iln A6 B24h A6 Inguinal - Skin Tail A6 RightEar - d-ln Right ear A6 LeftEar - Skin - d-ln - Skin Left ear Numer of T cells per 4ng of tissue Skin - d-ln A6 LeftEar A6 RightEar - Skin A6 Inguinal - d-ln - iln - Skin Tail - Blood A6 B24h Left ear Right ear Supplementary Fig. 2. Peripheral TCR presence is not due to tissue contamination y lood T cells. Nature Medicine: doi:.38/nm.386

12 a Seitization Day, Boost Day 7 Challenge Day mouse Seitized mice Skin (tail) Distant (tail) Challenged (R. ear) Se. + chall. (L. ear) CD4 CD8 c LN Draining LN CD4 CD8 d Ratio (to gamma-delta) CD4 CD8 p=.2 /.2 /.4 p=.9 p=.4 e Distant Challenged Se.+chall. Supplementary Fig. 3. Frequency of CD4 and CD8 T cells in local and distant skin sites and LN, after multiple immunization. Nature Medicine: doi:.38/nm.386

13 a CD44 CD62L CD69 CD3 CD22 CD T RM T CM CD8 ESL (Ca++) ESL (EDTA) PSL (Ca++) PSL (EDTA) T RM T CM CD8 Supplementary Fig. 4. Phenotype of CD8 T cells present in LN (T CM ) and skin (T RM ) days after immunization (memory phase). Nature Medicine: doi:.38/nm.386

14 a c Seitiz. Day -7 Day,,7 Day Tail skin / Left inguinal LN Paraiotic surgery, c 4 weeks Day 63 Tail skin, ear skin / cervical LN, right inguinal LN Ear Thickness (xum) NS Neg Day Day 8 Skin Lymph nodes d e Before paraiosis After paraiosis p=.44 Paraiont pairs Non-Paraiont pairs # of shared T cell clones 5 5 Tail skin Lymph node p=. Before paraiosis After paraiosis # of shared T cell clones 5 5 # H H3 H5 H7 H9 H Mice -6 H H9 H7 H5 H3 H f Ear skin 6 w post to ear Specific TCR! counts CASSDRGGYNSLYF Common sequences g CASSDRGGYNSLYF sequ. count - Se. Paraiont Paraiont NT - LN Tail LN Ear Tail Pre-Se. Post-Se. and Para. Tail skin 6w post to ear Activated naïve T cell clones give rise to oth skin-resident T RM and re- Supplementary Fig. 5. circulating T CM. Nature Medicine: doi:.38/nm.386

15 a Draining (axillary) LN vs distant (inguinal) LN Immunized (left) ear vs distant (right) ear TCR overlap (% of unique T cell clones) 5% % 5% p=.62 p=.3 % % 5% 7% 6% % % 5% TCR overlap (% of total T cells) TCR overlap (% of unique T cell clones) 5% % 5% p=.4 5% % p=. 5% % 6% % % 5% TCR overlap (% of total T cells) % % % % Seitized Clones 5 cells in L&R ear 2 24 Also in tail (5%) 24 (%) Also in aln (5%) 23 (96%) Also in iln ( 5%) 7 (7%) Mean clone size c T cell clone aundance in SKIN d T cell clone aundance in LN Proportion of unique T cell clones % % %.% 6 (89.6%) 59 (72.6%) 8 (22.5%) 7 (.%) 8 (2.2%) 22 (2.8%) 2 (.2%) (.%) Proportion of unique T cell clones 4 (97.3%) 3 5 (96.3%) % % %.% (3.5%) 38 (2.7%) 3 (.2%) 6 (.2%) 3 (.2%) 3 (.%).% cell 2-5 cells 6- cells +cells.% cell 2-5 cells 6- cells +cells e Proportion of unique T cell clones T cell clone aundance in LN, among those 5 in the skin % 9 (38%) 4 (7%) 5 (2%) 3.5 (5%) 2.5 (%) % %.% 6 week old / naïve mice weeks old / exp..% cell cell 2-5 cells 6- cells +cells Supplementary Fig. 6. TCR sequence overlap in lymph nodes and skin at different life-stages (naïve and antigen experienced). Nature Medicine: doi:.38/nm.386

16 42.5 PerCP-A PerCP-A a Day -,2 I II Seitization Day, Boost Day 7 Challenge Day +FTY7 Adoptive trafer (dln) /c d/e CD3 Ratio: FTY7/Untreate FTY c SSC SSC-A SSC-A CD CD8 CD4 Vehicle FTY7 III IV V VI VII +FTY7 +FTY7 +FTY7 +FTY7 +FTY7 +FTY7 f/g h/i j/k l/m n/o d f Ear Thickness (xum) Ear Thickness (xum) 4 +FTY7 AT AT + FTY Left ear 4 p=. Ear Thickness (xum) post adoptive trafer post adoptive trafer +FTY7 S+B / Vehicle 5 S+B + FTY (d) 4 p< e +FTY7 Right ear p=. 3 AT AT + FTY g Ear Thickness (xum) S+B / vehicle S+B / +FTY (d) D D8 D D3 h i Ear Thickness (xum) +FTY7 5 S+B / Se. ear S+B / ear S+B + FTY (f) / Se. ear 4 θ S+B + FTY (f) / ear p=. 2 p=. 3 θ p= Time to Max (days) Se. Se. Vehicle + FTY j Ear'Thickness'(xum) 5 4 Sei/zed2+2oosted Sei/zed2only2(h) p=. 6 p=. 7 p= k Time to Max (days) p=e-7 S+ oost S only l Ear Thickness (xum) +FTY7 p=. Setized only (h) Seitized only + FTY (j) m Max Ear Thickness (xum) p=. p=. (h) (j) n Ear Thickness (xum) +FTY7 5 Seitized + oost FTY during Se. + oost (l) 4 3 p=. 3 p= o Time to Max (days) S + B p=.2 (l) FTY d.s+b Supplementary Fig. 7. The fast memory respoe to is independent of T CM migration, while the slow memory respoe is T CM migration dependent. Nature Medicine: doi:.38/nm.386

17 a Seitization Day, Boost Day 7 Challenge Day Late chall. Day 6 Very late c. Day 9 I c d II Adoptive trafer (dln) e g c d Ear Thickness (xum) p<. p=. 2 5 Seitized () 4 4 Ear Thickness (xum) p=. 5 p= Se. (d) Se. (c) Ear Thickness (xum) e Ear Thickness (xum) p=.4 AT: Donor AT: Donor Se. (e) post adoptive trafer f Time to Max (days) p=.5 n.s. 6 9, c, d e g Ear Thickness (xum) AT: Seitiz.D7 AT: Donor Se. D p=. 4 p= post adoptive trafer h Time to Max (days) p=.4 D7 Tr. D Tr. Supplementary Fig. 8. T cell respoe kinetics of T RM and T CM are different yet oth are long lasting. Nature Medicine: doi:.38/nm.386

18 a c Ear Thickness (xum) Seitization Day, 5 4 Boost Day 7 Challenge Day "" ear (right) "Se." ear (left) p=.6 p=. 2nd chall. Day 6 3 rd chall. Day 9 c/d " p= Ear Thickness (xum) p=. 2 " d p=. 4 Ear Thickness (xum) 5 4 Start " p=. 3 Se. ear (S+B) ear (neither S nor B) "Se." ear (left) "" ear (right) " Boost (local) st Ch. 2nd Ch. 3rd Ch. p=. p=.6 " p=. Supplementary Fig. 9. Repeated antigenic skin exposure increases the inteity of the T RM respoe. Nature Medicine: doi:.38/nm.386

19 a Human Skin, patient B Unrelated human skin samples Human Skin, patient A, Placeo skin Human Skin, patient A, 3 days after Related human skin samples same day (+/- DPCP) Human Skin, patient A, Placeo skin Human Skin, patient A, 4 month after DPCP Related human skin samples long interval (4 month). Human Skin, patient A, Placeo skin D3 Placeo vs. D3 DPCP D DPCP vs. Day 4 DCPC c Quantification (Normalized TCR counts / 4ng tissue) Hu Skin, patient A, 3 days after DPCP a c -4 Human Skin, patient A, Placeo skin Hu Skin, 4 days after DPCP exp. 4 c 2 3 a. Human Skin, patient A, 4 month after DCPC exp. D3_Placeo D3_DPCP D4_DPCP D_DPCP Supplementary Fig.. TCR clone profiling during DPCP mediated contact dermatitis. Nature Medicine: doi:.38/nm.386