Decreased diversity and increased clonality of foetal compared to young adult TCRβ repertoires

TCR β-chain repertoires were sequenced from FACS-sorted CD4+CD8+ (DP), CD4+CD8-CD3+ (SP4) and CD4-CD8+CD3+ (SP8) thymocyte populations from E18.5 (red circles; n=7) and 4 week-old (blue circles; n=6) C57BL/6 thymus.

(A) The frequency distribution of TCR β-chain abundance was fitted to a discrete power law (f(k) = Ck−α) by maximum likelihood (solid line). These logged plots show a representative individual mouse TCR repertoire frequency distribution for E18.5 and 4 weeks in DP, SP4 and SP8 populations. The x axis represents TCR abundance (size of clone), and the y axis represents the proportion of the repertoire. The negative of the power law exponent corresponds to the slope of the logged plot. (B) The power law exponents of the frequency distribution of TCR β-chain abundances for E18.5 and 4 weeks in DP, SP4 and SP8 populations, where each point represents an individual mouse/embryo. (C)The proportion of the total TCR β-chain repertoire accounted for by the expanded top 1% most abundant sequences (>99th percentile) for E18.5 and 4 weeks for DP, SP4 and SP8 populations, where each point represents an individual mouse/embryo. (D) The number of β-chain sequences detected above the given frequency threshold (top 1% most abundant sequences) is shown for E18.5 (red circles) and 4 weeks (blue circles) in DP, SP4 and SP8 populations. Each small translucent point represents the abundance of any β-chain sequence detected above the threshold for all mice/embryos at that life-stage, while each larger solid point represents the mean of β-chain sequence abundance for each individual mouse/embryo. (E) The Shannon entropies for the β-chain TCR repertoires of E18.5 and 4 weeks in DP populations, where each point represents an individual mouse/embryo. Each repertoire was subsampled to 50000 TCRs 1000 times before calculating the Shannon entropy. (F) The Shannon entropies for the β-chain TCR repertoires of E18.5 and 4 weeks in SP4 and SP8 populations, where each point represents an individual mouse/embryo. Each repertoire was subsampled to 750 TCRs 1000 times before calculating the Shannon entropy. (G) The Gini indexes for the β-chain TCR repertoires of E18.5 and 4 weeks in DP populations, where each point represents an individual mouse/embryo. Each repertoire was subsampled to 50000 TCRs 1000 times before calculating the Gini index. (H) The Gini indexes for the β-chain TCR repertoires of E18.5 and 4 weeks in SP4 and SP8 populations, where each point represents an individual mouse/embryo. Each repertoire was subsampled to 750 TCRs 1000 times before calculating the Gini index.

Dotplots show mean±c.i and statistical comparisons were carried out by unpaired Student’s t-test or Welch’s t-test as appropriate: *** p <0.001; ** p <0.01; *p < 0.05.

Foetal and young adult TCR repertoires favour different gene segments

TCR α-chain (A-B) and β-chain (C-G) repertoires were sequenced from FACS-sorted CD4+CD8-CD3+ (SP4) and CD4-CD8+CD3+ (SP8) CD4+CD8+CD69- (CD69-DP, foetal), CD4+CD8+CD69+ (CD69+DP foetal) CD4+CD8+CD3-/lo (CD3-/loDP, adult), CD4+CD8+CD3+/hi (CD3+/hiDP adult) thymocyte populations from E18.5 (red; n=7) and 4 week-old (blue; n=6) C57BL/6 thymus. DP, SP4 and SP8 populations are coloured in green, orange and purple respectively.

(A-D) Heatmaps of proportional α-chain variable (V) (A), α-chain joining (J) (B), β-chain V(C) and β-chain J (D) gene usage of unique TCRs for E18.5 and 4 weeks in DP, SP4 and SP8 populations. Each column represents a mean of 7 embryos or 6 mice and was clustered using Euclidian distance. Genes are shown in chromosomal order (5’ to 3’) from top to bottom. (E-G) PCA biplot of β-chain V and J gene counts of unique TCRs for E18.5 and 4 weeks in DP, (E) SP4 (F) and SP8 (G) populations.

Principal Component Analysis (PCA) of Variable x Joining combinations in embryonic and young adult TCRβ repertoires cluster by life-stage before cell-type

TCR β-chain repertoires were sequenced from FACS-sorted CD4+CD8+ (DP), CD4+CD8-CD3+ (SP4) and CD4-CD8+CD3+ (SP8) thymocyte populations from E18.5 (n=7) and 4 week-old (n=6) C57BL/6 thymus.

(A-B) PCA biplot of β-chain VxJ gene counts of unique TCRβ sequences for E18.5 and 4 weeks in DP, SP4 and SP8 populations. (A) E18.5 and 4 weeks are coloured in red and blue circles respectively. (B) DP, SP4 and SP8 populations are coloured in green, orange and purple respectively. (C) Top 20 highest β-chain VxJ gene combinations that contribute to PC1 and PC2 of the PCA.

Proportional bias in VxJ combinations in foetal and young adult repertoires by cell-type

TCR β-chain and α-chain repertoires were sequenced from FACS-sorted CD4+CD8+ (DP), CD4+CD8-CD3+ (SP4) and CD4-CD8+CD3+ (SP8) thymocyte populations from E18.5 (red; n=7) and 4 week-old (blue; n=6) C57BL/6 thymus.

(A-C) Plots show β-chain (left column) and α-chain (right column) proportional VxJ gene usage of total TCRs for E18.5 compared to 4 weeks in DP (A), SP4 (B) and SP8 (C) populations. Red tiles signify increased usage of VxJ combinations in E18.5 (p < 0.05), while blue tiles signify decreased usage in E18.5 (p < 0.05) compared to 4 weeks. Light grey tiles signify no change in gene usage (p > 0.05), yellow tiles signify VxJ combinations not detected in E18.5, brown tiles signify VxJ combinations not detected in 4 weeks and black tiles signify VxJ combinations not detected in both E18.5 and 4 weeks. Only combinations that were detected in at least 3 mice mice/embryos per group were compared. Dark grey tiles signify combinations that were not compared. Statistical comparisons were carried out by unpaired Student’s t-test followed by FDR-adjustment (5%, Benjamini-Hochberg procedure) of p values.

Decreased TCR β non-template insertions, CDR3 length with increased CDR3 sharing, but reduced amino acid motif sharing in foetal compared to young adult

TCR β-chain repertoires were sequenced from FACS-sorted CD4+CD8+ (DP), CD4+CD8-CD3+ (SP4) and CD4-CD8+CD3+ (SP8) thymocyte populations from E18.5 (red circles; n=7) and 4 week-old (blue circles; n=6) C57BL/6 thymus. (A) Percentage of non-productive β-chain TCRs from total rearrangements for E18.5 and 4 weeks in DP, SP4 and SP8 populations. (B) The weighted mean of non-template insertion length for all β -chain TCRs (basepairs) for E18.5 and 4 weeks in DP, SP4 and SP8 populations. (C) The mean of β-chain unique predicted CDR3 length (number of amino acids) for E18.5 and 4 weeks in DP, SP4 and SP8 populations. (D) The intra life-stage Jaccard Index of Similarity of β-chain CDR3s for E18.5 and 4 weeks in the DP population. Each repertoire was subsampled to 30000 CDR3s 1000 times before calculating the Jaccard Index. (E) The intra life-stage Jaccard Index of Similarity of β-chain CDR3s for E18.5 and 4 weeks in the SP4 and SP8 population. Each repertoire was subsampled to 500 CDR3s 1000 times before calculating the Jaccard Index. Dotplots show mean±c.i and statistical comparisons were carried out by unpaired Student’s t-test or Welch’s t-test as appropriate: *** p <0.001; ** p <0.01; *p < 0.05.

Proportional combinatorial CDR1xCDR2 bias in foetal and young adult thymocyte populations indicates reduced MHC-restriction in foetal thymus

TCR β-chain and α-chain repertoires were sequenced from FACS-sorted CD4+CD8+ (DP), CD4+CD8-CD3+ (SP4) and CD4-CD8+CD3+ (SP8) thymocyte populations from E18.5 (n=7) and 4-week-old (n=6) C57BL/6 thymus and frequency of CDR1 and CDR2 sequences calculated for each population. E18.5 and 4-week samples are coloured red and blue respectively, while DP, SP4 and SP8 populations are coloured green, orange and purple respectively. (A-B) PCA biplot of β-chain CDR1xCDR2 frequency distributions for E18.5 and 4 weeks in DP, SP4 and SP8 populations. In (A), E18.5 and 4-weeks samples are coloured in red and blue respectively, while in (B), DP, SP4 and SP8 populations are coloured in green, orange and purple respectively. (C) Top 10 highest CDR1xCDR2 combinations that contribute to PC1 and PC2 of the PCA of β-chain CDR1xCDR2 frequency distributions for E18.5 and 4 weeks in DP, SP4 and SP8 population (shown in A-B). (D) Heatmap of proportional β-chain CDR1xCDR2 usage E18.5 and 4 weeks in DP, SP4 and SP8 populations. Each column represents a mean of 7 embryos or 6 adult mice and was clustered using Euclidian distance, while rows (CDR1 x CDR2 combinations) were clustered using Pearson correlation. (E) Heatmap of proportional α-chain CDR1xCDR2 usage in E18.5 and 4 week samples in DP, SP4 and SP8 populations. Each column represents a mean of 7 embryos or 6 mice and was clustered using Euclidian distance, while rows (CDR1xCDR2 combinations) were clustered using Pearson correlation.

PCA of VxJ and CDR1xCDR2 combinations in foetal and young adult thymocyte populations indicate reduced MHC-restriction of foetal TCR repertoires

TCR β-chain and α-chain repertoires were sequenced from FACS-sorted CD4+CD8+ (DP), CD4+CD8-CD3+ (SP4) and CD4-CD8+CD3+ (SP8) thymocyte populations from E18.5 (n=7) and 4 week-old (n=6) C57BL/6 thymus.

(A-B) PCA biplot of β-chain VxJ gene counts of unique TCRβ sequences for E18.5 (A) and 4 week (B) in DP, SP4 and SP8 populations (coloured in green, orange and purple respectively). The percentage of variance for each principal component is indicated on the axis. (C) Ten β-chain VxJ gene combinations that contribute most to PC1 and PC2 of the PCA for E18.5. (D) Ten β-chain VxJ gene combinations that contribute most to PC1 and PC2 of the PCA for 4 weeks. (E-F) PCA biplot of α-chain VxJ gene counts of unique TCRβ sequences for E18.5 (E) and 4 weeks (F) in DP, SP4 and SP8 populations (coloured in green, orange and purple respectively). The percentage of variance for each principal component is indicated on the axis. (G-H) PCA biplot of β-chain CDR1xCDR2 frequency distributions of unique TCRβ sequences for E18.5 (G) and 4 weeks (H) in DP, SP4 and SP8 populations (coloured in green, orange and purple respectively). The percentage of variance for each principal component is indicated on the axis. (I) Ten β-chain CDR1xCDR2 combinations that contribute most to PC1 and PC2 of the PCA of β-chain CDR1xCDR2 frequency distributions for E18.5. (J) Ten β-chain CDR1xCDR2 combinations that contribute most to PC1 and PC2 of the PCA of β-chain CDR1xCDR2 frequency distributions for 4 weeks. (K-L) PCA biplot of α-chain CDR1xCDR2 frequency distributions of unique TCRβ sequences for E18.5 (K) and 4 weeks (L) in DP, SP4 and SP8 populations (coloured in green, orange and purple respectively). The percentage of variance for each principal component is indicated on the axis.

In hydrocortisone-depleted adult thymus recovering DP populations use foetal-like gene segments

TCR α-chain (A-C) and β-chain (D-F) repertoires were sequenced from FACS-sorted CD4+CD8+CD69- (CD69-DP, foetal), CD4+CD8+CD69+ (CD69+DP foetal) CD4+CD8+CD3-/lo (CD3-/loDP, adult), CD4+CD8+CD3+/hi (CD3+/hiDP adult) thymocyte populations from E18.5 (red; n=7), 4 week-old (blue; n=6), and 4 week-old treated with hydrocortisone on day 6 after treatment (light blue; n=5) C57BL/6 thymus.

(A-B) Heatmaps of proportional α-chain variable (V) (A) and α-chain joining (J) (B) gene usage of productive TCRs for E18.5, 4 weeks (control) and 4 weeks treated with hydrocortisone in DP populations. Each column represents a mean of 7 or 6 (E18.5 and 4 weeks) or 5 (4 weeks treated with hydrocortisone) embryos or mice and was clustered using Euclidian distance. Genes are shown in chromosomal order (5’ to 3’) from top to bottom. (C) Plot shows α-chain proportional VxJ gene usage of total TCRs for 4 weeks HC-treated compared to 4 weeks (control) in DP populations. Pink tiles signify increased usage of VxJ combinations in 4 weeks HC-treated (p < 0.05), while green tiles signify decreased usage in 4 weeks HC-treated (p < 0.05) compared to 4 weeks (control). Grey tiles signify no change in gene usage (p > 0.05), yellow tiles signify VxJ combinations not detected in 4 weeks HC-treated, brown tiles signify VxJ combinations not detected in 4 weeks (control) and black tiles signify VxJ combinations not detected in both 4 weeks HC-treated and 4 weeks (control). Only combinations that were detected in at least 3 mice per group were compared. Dark grey tiles signify combinations that were not compared. Statistical comparisons were carried out by unpaired Student’s t-test followed by FDR-adjustment (5%, Benjamini-Hochberg procedure) of p values. (D-E) Heatmaps of proportional β-chain V(D) and β-chain J (E) gene usage of unique TCRs for E18.5, 4 weeks (control) and 4 weeks treated with hydrocortisone in DP populations. Each column represents a mean of 7 or 6 (E18.5 and 4 weeks) or 5 (4 weeks treated with hydrocortisone) embryos or mice and was clustered using Euclidian distance. Genes are shown in chromosomal order (5’ to 3’) from top to bottom. (F) Plot shows β-chain proportional VxJ gene usage of total TCRs for 4 weeks HC-treated compared to 4 weeks (control) in DP populations. Pink tiles signify increased usage of VxJ combinations in 4 weeks HC-treated (p < 0.05), while green tiles signify decreased usage in 4 weeks HC-treated (p < 0.05) compared to 4 weeks (control). Grey tiles signify no significant change in gene usage (p > 0.05), yellow tiles signify VxJ combinations not detected in 4 weeks HC-treated, brown tiles signify VxJ combinations not detected in 4 weeks (control) and black tiles signify VxJ combinations not detected in both 4 weeks HC-treated and 4 weeks (control). Statistical comparisons were carried out by unpaired Student’s t-test followed by FDR-adjustment (5%, Benjamini-Hochberg procedure) of p values.

Distribution and diversity of embryonic and young adult TCRα repertoire

TCR α-chain repertoires were sequenced from FACS-sorted CD4+CD8+ (DP), CD4+CD8-CD3+ (SP4) and CD4-CD8+CD3+ (SP8) thymocyte populations from E18.5 (red circles; n=7) and 4 week-old (blue circles; n=6) C57BL/6 thymus. (A) The frequency distribution of TCR α-chain abundance was fitted to a discrete power law (f(k) = Ck−α) by maximum likelihood. These logged plots show a representative individual mouse TCR repertoire frequency distribution for E18.5 and 4 weeks in DP, SP4 and SP8 populations. The x axis represents TCR abundance (size of clone), and the y axis represents the proportion of the repertoire. The negative of the power law exponent corresponds to the slope of the logged plot. (B) The power law exponents of the frequency distribution of TCR α-chain abundances for E18.5 and 4 weeks in DP, SP4 and SP8 populations, where each point represents an individual mouse/embryo. (C) The proportion of the total TCR α-chain repertoire accounted for by the expanded top 1% most abundant sequences (>99th percentile) for E18.5 and 4 weeks for DP, SP4 and SP8 populations, where each point represents an individual mouse/embryo. (D) The number of α-chain sequences detected above the given frequency threshold (top 1% most abundant sequences ) is shown for E18.5 (red circles) and 4 weeks (blue circles) in DP, SP4 and SP8 populations. Each small translucent point represents the abundance of any β-chain sequence detected above the threshold for all mice/embryos at that life-stage, while each larger solid point represents the mean of α-chain sequence abundance for each individual mouse/embryo. (E) The Shannon entropies for the α-chain TCR repertoires of E18.5 and 4 weeks in DP populations, where each point represents an individual mouse/embryo. Each repertoire was subsampled to 50000 TCRs 1000 times before calculating the Shannon entropy. (F) The Shannon entropies for the α-chain TCR repertoires of E18.5 and 4 weeks in SP4 and SP8 populations, where each point represents an individual mouse/embryo. Each repertoire was subsampled to 750 TCRs 1000 times before calculating the Shannon entropy. (G) The Gini indexes for the α-chain TCR repertoires of E18.5 and 4 weeks in DP populations, where each point represents an individual mouse/embryo. Each repertoire was subsampled to 50000 TCRs 1000 times before calculating the Gini index. (H) The Gini indexes for the α-chain TCR repertoires of E18.5 and 4 weeks in SP4 and SP8 populations, where each point represents an individual mouse/embryo. Each repertoire was subsampled to 750 TCRs 1000 times before calculating the Gini index.

Dotplots show mean±c.i and statistical comparisons were carried out by unpaired Student’s t-test or Welch’s t-test as appropriate: *** p <0.001; ** p <0.01; *p < 0.05.

Foetal and young adult DP populations have proportionally different TRAV, TRAJ, TRBV and TRBJ gene usage

(A-D) Dotplots show proportional TRAV (A), TRAJ (B), TRBV (C), and TRBJ (D) gene usage of unique TCRs for E18.5 (red circles, n=7) and 4 weeks (blue circles, n=6) in DP populations. Genes are shown in chromosomal order (5’ to 3’) from left to right. Dots represent individual mice (n=7 or 6) and bars show the mean. Statistical comparisons were carried out by unpaired Student’s t-test or Welch’s t-test as appropriate. Genes significantly increased were highlighted in red (when increased in E18.5) or blue (when increased in 4 weeks). *** p <0.001; ** p <0.01; *p < 0.05.

Foetal and young adult SP4 populations have proportionally different TRAV, TRAJ, TRBV and TRBJ gene usage

(A-D) Dotplots show proportional TRAV (A), TRAJ (B), TRBV (C), and TRBJ (D) gene usage of unique TCRs for E18.5 (red circles, n=7) and 4 weeks (blue circles, n=6) in SP4 populations. Genes are shown in chromosomal order (5’ to 3’) from left to right. Dots represent individual mice and bars show the mean. Statistical comparisons were carried out by unpaired Student’s t-test or Welch’s t-test as appropriate. Genes significantly increased were highlighted in red (when increased in E18.5) or blue (when increased in 4 weeks). *** p <0.001; ** p <0.01; *p < 0.05.

Foetal and young adult SP8 populations have proportionally different TRAV, TRAJ, TRBV and TRBJ gene usage

(A-D) Dotplots show proportional TRAV (A), TRAJ (B), TRBV (C), and TRBJ (D) gene usage of unique TCRs for E18.5 (red circles, n=7) and 4 weeks (blue circles, n=6) in SP8 populations. Genes are shown in chromosomal order (5’ to 3’) from left to right. Dots represent individual mice (n=4) and bars show the mean. Statistical comparisons were carried out by unpaired Student’s t-test or Welch’s t-test as appropriate. Genes significantly increased were highlighted in red (when increased in E18.5) or blue (when increased in 4 weeks). *** p <0.001; ** p <0.01; *p < 0.05.

Proportional TRBJ gene segment usage from TRBJ1 and TRBJ2 gene clusters

(A-B) Dotplots show proportional TRBJ gene segment usage from TRBJ1 cluster (A) and TRBJ2 cluster (B), for E18.5 (red circles, n=7) and 4 weeks (blue circles, n=6) in DP, SP4 and SP8 populations. Statistical comparisons were carried out by unpaired Student’s t-test. *** p <0.001; ** p <0.01; *p < 0.05.

TCR α CDR3 non-template insertions, length, sharing and PCA of α-chain CDR1xCDR2 frequency distributions in foetal and young adult thymocyte populations

TCR α-chain repertoires were sequenced from FACS-sorted CD4+CD8+ (DP), CD4+CD8-CD3+ (SP4) and CD4-CD8+CD3+ (SP8) thymocyte populations from E18.5 (red circles; n=7) and 4 week-old (blue circles; n=6) C57BL/6 thymus.

(A) Percentage of non-productive α-chain TCRs from total rearrangements for E18.5 and 4 weeks in DP, SP4 and SP8 populations. (B) The weighted mean of non-template insertion length for all α-chain TCRs (basepairs) for E18.5 and 4 weeks in DP, SP4 and SP8 populations. (C) The mean of α-chain unique predicted CDR3 length (number of amino acids) for E18.5 and 4 weeks in DP, SP4 and SP8 populations. (D) The intra life-stage Jaccard Index of Similarity of α-chain CDR3s for E18.5 and 4 weeks in the DP population. Each repertoire was subsampled to 30000 CDR3s 1000 times before calculating the Jaccard Index. (E) The intra life-stage Jaccard Index of Similarity of α-chain CDR3s for E18.5 and 4 weeks in the SP4 and SP8 population. Each repertoire was subsampled to 500 CDR3s 1000 times before calculating the Jaccard Index.

(F-G) PCA biplot of α-chain CDR1xCDR2 frequency distributions for E18.5 and 4-weeks in DP, SP4 and SP8 populations. In (F), E18.5 and 4-weeks samples are coloured in red and blue respectively, while in (G), DP, SP4 and SP8 populations are coloured in green, orange and purple respectively. Dotplots show mean±c.i and statistical comparisons were carried out by unpaired Student’s t-test or Welch’s t-test as appropriate: *** p <0.001; ** p <0.01; *p < 0.05.

Gating strategy for cell sorting for foetal thymus.

Figure shows representative FACS plots and gates for cell sorting of foetal thymus.

Gating strategy for cell sorting for adult thymus.

Figure shows representative FACS plots and gates for cell sorting of adult thymus.