Basic information of the subjects

A total of 13 subjects in the IOI group and 6 subjects in the HC group were included in this study. Before treatment, blood samples were collected from all subjects in the IOI and HC group. Unfortunately, after the treatment of glucocorticoids, we only received blood samples of 6 patients in IOI group.

A comparison of the basic information of the IOI and HC groups is shown in Table 1, and the differences in BCVA and D50 between the two groups were statistically significant. The mean follow-up and scoring time of the 13 patients in the IOI group was 6.69 ± 4.03 months. The patients in the IOI group were scored for disease activity before and after treatment according to the modified Werner grading method [24] (Supplemental Table S1). Figure 1 shows the comparison of clinical examinations in the IOI(EF) group and the IOI(IN) group.

Table 1 Basic information of the subjectsFig. 1

Comparison of clinical examinations in the IOI(EF) group and the IOI(IN) group. (A. B): MRI of IOI03 in the IOI(EF) group showed swelling of the right eye extraocular muscle before treatment. (C. D): After treatment, MRI showed obvious improvement in the above symptoms of IOI03. (E. F): CT of IOI07 in the IOI(IN) group showed swelling of the extraocular muscle of the right eye. (G. H): CT of IOI07 after treatment showed no improvement of the above symptoms. (I): IOI01 in IOI(EF) group showed obvious swelling of the upper eyelid of the right eye before treatment. (J): After treatment, the above symptoms of IOI01 were significantly improved. (K): IOI07 in the IOI(IN) group showed ptosis of the upper eyelid and swelling of right eye before treatment. (L): The above symptoms of IOI07 did not improve after treatment

TCR CDR3 sequence and data analysis

The peripheral blood TCR profiles of all subjects were obtained by high-throughput sequencing. After quality control, we obtained the mean TCR-CDR3 clean reads for each sample in the IOI group as 102,102,541 and the mean TCR-CDR3 clean reads for each sample in the HC group as 89,791,638. The minimum value in the IOI group was 82,737,122 and the maximum value in the IOI group was 113,676,620. The minimum value in the HC group was 72,502,404 and the maximum value in the HC group was 132,772,068. Functional analysis of the TCR rearrangement sequences of each sample revealed that more than 85% of the TCR-encoded products with functions were found in each sample. (Supplemental Table S2).

CDR3 polypeptide chain length distribution characteristics and TRBV and J gene usage

The diversity of TCR CDR3 repertoire was correlated with the recombination V, D, and J genes [29]. During recombination, exonuclease sheared the end of the recombinant gene as well as terminal deoxynucleotidyl transferase (TdT) introduced additional nucleotides at the D-J and V-D junctions of the recombinant gene. These processes caused changes in the length of the CDR3 [30, 31]. Therefore, we calculated the cumulative frequencies of V, D, and J structural domains and V-J paired gene use types and CDR3 lengths. We found that the distribution of CDR3 length in the IOI group was similar to that of the HC group, mostly between 13 and 16 nt, with a peak accumulation frequency of 15 nt (Fig. 2).

Fig. 2

Length distribution of TCR CDR3 in the IOI and the HC groups

The length distribution of TCR CDR3 amino acid of the IOI group (n = 13) and HC group (n = 6). The amino acid length showed a peak at 15 nt of both two groups.

Although the type of TRBV gene use was essentially the same in the IOI and HC groups, we found significant differences in the frequency of partial gene accumulation in the two groups. Among them, the frequencies of TRBV5-1, TRBV6-2, TRBV6-5, and TRBV6-6 use were significantly lower in the IOI group than in the HC group (P = 0.010, P = 0.019, P = 0.024, P = 0.011) (Fig. 3A).

At the TRBV20-1 and TRBV29-1 sequences, the frequency of use showed a trend of difference that was not significant between the IOI and the HC groups (P = 0.065, P = 0.065) (Fig. 3A). Among the 13 IOI samples before glucocorticoid therapy, the type of V gene use was the same between the IOI(EF) and the IOI(IN) groups, and the frequency of use at TRBV4-2 and TRBV6-6 was significantly higher in the IOI(IN) group than in the IOI(EF) group (P = 0.041, P = 0.032) (Fig. 3C).

In the analysis of the frequency of J gene use, there was no significant difference between the IOI and HC groups (Fig. 3B). However, between the IOI(IN) group and the IOI(EF) group, there were statistically significant differences in the use of sample genes on J1-2, J1-5 and J1-6 (P = 0.005, P = 0.003, P = 0.026) (Fig. 3D).

Fig. 3

Usage of the V、J segment in the different groups. Frequency distribution of TRBV gene (A) and TRBJ gene (B) in the IOI group (n = 13) and the HC group (n = 6). Frequency distribution of TRBV gene (C) and TRBJ gene (D) in the IOI(EF) group (n = 6) and the IOI(IN) group (n = 7)

TCR CDR3 diversity in the samples

To quantify the TCR diversity in the IOI and HC groups, we used Shannon entropy as well as D50.

We analyzed the TCR CDR3 sequences of each group and found that both Shannon entropy and D50 were significantly lower in the IOI group than in the HC group (P < 0.001, P = 0.002). Besides, Shannon entropy and D50 were significantly lower in the IOI(EF) group than in the HC group (P < 0.001, P < 0.001). Shannon entropy and D50 were significantly lower in the IOI(IN) group than in the HC group (P < 0.001, P = 0.003) (Fig. 4A, B). Next, we compared the diversity of the IOI(EF) and the IOI(IN) group before glucocorticoid treatment. Shannon entropy showed a trend of difference that was not significant between the IOI(EF) group and the IOI(IN) group (P = 0.053) (Fig. 4A). Finally, D50 was significantly lower in the IOI(EF) group than in the IOI(IN) group (P = 0.038) (Fig. 4B).

Fig. 4

Comparison of diversity among groups (IOI, HC, IOI(EF) and IOI(IN)). Comparison of Shannon entropy (A) and D50 (B) among groups. The IOI group (n = 13), the HC group (n = 6), the IOI(EF) group (n = 6) and the IOI(IN) group (n = 7)

Blood samples were also collected from six IOI patients after glucocorticoid therapy, including four in the IOI(EF) group and two in the IOI(IN) group. By comparing the Shannon entropy before and after treatment, it was found that the diversity of these six samples showed an increasing tendency that was not significant after treatment(P = 0.063). (Supplemental Table S3).

Patterns of TCR CDR3 sequence sharing among individuals

As mentioned above, the degree of clonal amplification and the frequency of gene accumulation differed between the IOI and HC groups, so it would be interesting to know if there is an overlap between the sequences of different individuals in each group. CDR3 is the most diverse region in the TCR involved in epitope recognition, so our analysis focused on the DNA and amino acid sequences of CDR3.

Notably, we found 133 CDR3 DNA sequences shared among the 13 IOI patients; these corresponded to 131 amino acid sequences, two of which had a frequency of more than 0.01% in each sample. In contrast, there were only three shared nucleotide sequences in the HC group, corresponding to three amino acid sequences. The 133 sequences shared by the IOI group were visible in 0 to 2 samples in the HC group, all with frequencies below 0.01%. Table 2 lists the top 10 nucleotide sequences and their corresponding amino acid sequences in the frequency of the shared sequences among the 13 IOI samples. We show the V-J gene pairs corresponding to these 10 sequences (Supplemental Table S4). The top 10 nucleotide sequences and their corresponding amino acid sequences are visible in some of the HC group samples, but the sequence percentage is extremely low. In addition, 133 CDR3 DNA sequences were shared among all IOI(EF) group samples, and the top 10 sequences with the highest frequencies were all visible in the IOI(IN) group. There were 175 CDR3 DNA sequences shared among all IOI(IN) group samples, and the top 10 sequences with the highest frequencies were all seen in the IOI(EF) group.

Table 2 The top 10 frequency sequences in the IOI sample-sharing sequence