TYK2 signaling promotes the development of autoreactive CD8+ cytotoxic T lymphocytes and type 1 diabetes

Tyrosine kinase 2 (TYK2), a member of the JAK family, has attracted attention as a potential therapeutic target for autoimmune diseases. However, the role of TYK2 in CD8+ T cells and autoimmune type 1 diabetes (T1D) is poorly understood. In this study, we generate Tyk2 gene knockout non-obese diabetes (NOD) mice and demonstrate that the loss of Tyk2 inhibits the development of autoreactive CD8+ T-BET+ cytotoxic T lymphocytes (CTLs) by impairing IL-12 signaling in CD8+ T cells and the CD8+ resident dendritic cell-driven cross-priming of CTLs in the pancreatic lymph node (PLN). Tyk2-deficient CTLs display reduced cytotoxicity. Increased inflammatory responses in β-cells with aging are dampened by Tyk2 deficiency. Furthermore, treatment with BMS-986165, a selective TYK2 inhibitor, inhibits the expansion of T-BET+ CTLs, inflammation in β-cells and the onset of autoimmune T1D in NOD mice. Thus, our study reveals the diverse roles of TYK2 in driving the pathogenesis of T1D.

expression.Further, although H2-K appears to be the more highly expressed MHC class I molecule in NOD mice, it would be interesting to know about the expression of H-2D, the 2nd major MHC class I molecule in NOD mice.This also applies to Fig.S4 F and G.   5).In the context of the lower expression of Fas and H2-K, what happened to beta cell function, such as insulin production?Was this also changed in the absence of Tyk2?This question could shed light on whether beta cells are more self-protected by down regulating molecules related to autoimmune attack or the beta cells may also down regulate production of self antigen(s).6).Fig. 4A the authors showed +/-24w-/-, what about -/-24w+/-?The authors also showed +/--/-10 days data, what about -/-+/-10 day results?This also applies to Fig. S4B.7) Fig. 4B, the authors showed that the proportion of IGRP tetramer positive CD8+ T cells remained stable between 6-wk-old and 11-wk-old mice.Based on the data, the author drew the conclusion that there was no epitope spreading.This reviewer does not think these data support the conclusion, as epitope spreading does not mean the alteration of a tested and disease-causing epitope.8).The authors showed CD44hi T cells, in most if not all, data, and the status of CD62L of the CD44hi T cells was not clear.It is also not clearly stated why the authors focused on CD44hi T cell populations.9).What was the gating used in Fig. 5E? 10).Based on some reduction of CXCR3+ CD8+ T cells in PLN but not iLN and spleen (Fig. 5F, G), the authors thought those T cells exited PLN and migrated to islets.This would be supportede if the authors showed the islet CD8+ T cells are CXCR3+.11).It would be important to test cross priming in the absence of IGRP peptide as the assay will test direct antigen presentation in the presence of peptide, but not cross presentation.The authors should test NY8.3T cells in response to NIT-1 cells in the presence of DCs but in absence of IGRP peptide.Alternatively, the authors need to pulse DCs with NIT-1 cells and use the pulsed DCs as antigen presenting cells to elicit NY8.3 T cell responses.12).The authors used 10 g/ml IGPR peptide in their assays; this concentration of peptide is extremely high for NY8.3 CD8+ T cells.13).The authors showed that Tyk2-/-IGRP CD8+ T cells had upregulated apoptosis-related genes including Casp3 and Birc5 (Fig. 5I), it is not clear if those cells are indeed prone to apoptosis.Annexin V staining would provide some information and/or explanation as to why the cells upregulated apoptosisrelated genes.14).The earliest time when the wild type (or +/-) NOD mice developed T1D was around 14 to 15-wks-old (Fig. 1A), whereas the earliest age when Tyk2-/-NOD mice developed T1D was ~22-wk-old (Fig. 1A).Early (from 6-wk-old) treatment of Tyk2 inhibitor BMS-986165 in wild type NOD mice showed further delay in T1D onset, ~32-wk-old, and marked reduced T1D incidence, ~20% (Fig. 6F), which is similar to the Tyk2-/mice shown in Fig. 1A.Tyk2 inhibitor treatment at pre-diabetic period (from 12 -wk-old) also strikingly delayed T1D onset (the first mouse developed diabetes was at 30-wk-old, Fig. 6G).This reviewer assumes that if the group size were larger, especially the treatment group, which had only 4 mice, the statistical outcome could be significant, even though the overall incidence of diabetes was not reduced.15).Many figure legends did not provide brief description about how the experiments were conducted.In this reviewer's opinion, the figure legend should be more informative, and the readers should not need to go to and from the main text and/or Materials/Methods to find the information about the figures.Further, it is not clear why the authors used MFI as the readout for most, if not all, the flow cytometric analysis.
Reviewer #2 (expert in JAK-STAT signalling and TYK2): The manuscript adds important insight and novelty in the cellular mechanisms in TYK2-mediated autoimmune type 1 diabetes and its potential prevention by a TYK2inib in preclinical mouse models.
Overall, the impressive data sets support the conclusions and are appropriately presented.Previously published, mostly the author's claims supporting work should be included and discussed appropriately (see comments) The methodology is state-of-the-art, provided details allow for reproducibility (but see comments) Major comments: • The authors should include in their introduction/discussion the relevant findings of Chandra et al (https://doi.org/10.1038/s41467-022-34069-z) in a human TYK2-deficient cell system • Fig2C the authors should consider to complement the heat map by an alternative visualisation of the crucial DEG and indicate in the results text more clearly which DEG (esp.IFNalpha and IFNbeta) might be of biological relevance albeit not reaching the applied statistical significance criteria.In this context the authors should also mention previous supportive work on human pancreatic beta-cells producing IFNalpa and CXCL10 under challenge conditions in a TYK2-dependent manner (Marroqui et al, https://doi.org/10.2337/db15-0362) • line 337ff, rationale for next experiments: the authors should acknowledge that impaired in vitro and in vivo killing capacity of TYK2-deficient CTLs has been already described (e.g.Simma et al, https://doi.org/10.1158/0008-5472.CAN-08-1705).
• line 477ff, discussion on potential risks/side effects of TYK2inib treatment: the authors should also mention the involvement of TYK2 in tumour surveillance and the potential cancer risk upon long-term deucravacitinib exposure (e.g.Yarmolinsky et al., https://doi.org/10.1002/ijc.34180).
• Technical/methodological details necessary for the understanding of the presented data a scattered rather randomly in the manuscript; e.g.insulitis scores 0-3 are defined in the legend to Figure S1 and not given in Material and Methods ‚Histology'; statistical stringency for RNASeq DEG is given in Material and Methods only and would help the reader if also mentioned in the results text or given in the legends of the main figures.
The authors are asked to revised the entire manuscript and indicate the necessary technical details in a clear and coherent manner.

Minor comments:
Line 129-30: The journal strongly encourages researchers to follow the 'Sex and Gender Equity in Research -SAGER -guidelines' and to include sex and gender considerations for studies involving humans, vertebrate animals and cell lines where relevant to the topic of study (an overview can be found here).The authors should comment on their findings.We wish to thank the Reviewer for your insightful comments, which have greatly helped us to improve the quality of our paper.In particular, the amount of IGRP peptide is critical for our experiments.
Thank you very much for pointing this out.The Reviewer's comments are written in blue.Our point-by-point responses to the Reviewer's comments are shown below.
Major: To investigate the impaired functions of CD8+ T cells in the absence of Tyk2, the authors generated Tyk2 deficient NY8.3 CD8 TCR transgenic NOD mice.The authors used Tyk2 deficient NY8.3 CD8+ T cells for some in vitro cytotoxicity and in vivo proliferation assays.However, it is important to know if diabetes development would be delayed and/or reduced in Tyk2 deficient NY8.3 NOD mice, especially as the authors found impaired function(s) of CD8+ T cells in the absence of Tyk2.
The Tyk2 deficient NY8.3 NOD mouse is the perfect tool to confirm and/or validate the authors' finding and the authors have these mice in hand.
We have added the following text in the Results and discussed these results.
Minor but not trivial: 1).T1D genetic susceptibility loci are more than those the author have detected using PCR.Although this reviewer thinks that the authors' Tyk2-/-NOD mice are likely to be mostly on NOD genetic background, a more rigorous test would be to carry out the genome scan by SNP linkage mapping.This can be done easily by various service providers, one of which is DartMouseTM (https://geiselmed.dartmouth.edu/dartmouse/).
Response: The Reviewer's comment is correct.We mistakenly described the results of the PCR analysis of T1D susceptibility loci as "all T1D genetic susceptibility loci…".This error has been corrected in accordance with the comment (as shown below).
We have revised the Results as described below.
(Line 125-129) "We analyzed 64 short tandem repeat (STR) loci in Tyk2KO.NOD mice and confirmed that all tested loci except for D9Mit83, which is located in chromosome 9 that has Tyk2 gene, were of NOD origin (Supplementary Fig. 1A).In addition, we confirmed STRs at the insulin-dependent diabetes susceptibility (Idd) loci (Idd1 to Idd15) were of NOD origin (Supplementary Fig. 2B) 28 ."Revised Supplementary Fig. 1A.The analysis of 64 short tandem repeat (STR) loci in female Tyk2KO.NOD mice.
3).Fig. 1C.What about the IAA in 24-week-old +/+ and +/-mice?Without the data from +/+ or +/-mice at a similar age, it is difficult to interpret the results that showed that IAA was significantly increased in -/-mice.The authors' interpretation related to aging is somewhat simplistic.
Revised Fig. 1B.Percentage of islets with a given insulitis score.
P-values were calculated using the one-way ANOVA with Tukey's posttest.
Revised Fig. 1C.Levels of serum IAAs.P-values were calculated using the one-way ANOVA with Tukey's posttest.
On the basis of these results, we have changed the following text from: "We also analyzed 24w mice to assess the long-term effects of Tyk2 deficiency.Increased numbers of inflamed islets and elevated levels of serum IAAs were observed in 24w Tyk2 -/-mice compared with 14w Tyk2 -/-mice (Fig. 1B, C).These observations suggest that Tyk2 deficiency does not completely prevent islet autoimmunity but reduces the progression rate of invasive insulitis leading to T1D onset." to (line 144-151) "We also analyzed normoglycemic 24w mice to assess the long-term effects of Tyk2 deficiency.Increased numbers of inflamed islets and elevated levels of serum IAAs were observed in 24w Tyk2 -/-mice compared with 14w Tyk2 -/-mice (Fig. 1B, C).Because diabetic mice were excluded, levels of insulitis and IAAs were understated in this analysis, especially in 24w Tyk2 +/+ and Tyk2 +/-mice.These observations suggest that Tyk2 deficiency does not completely prevent islet autoimmunity but reduces the progression rate of invasive insulitis leading to T1D onset."4).On line 194-195 of the main text, the authors stated that MHC class I, H-2K, was reduced in expression on islet beta cells whereas B2m was more highly expressed.The authors did not show B2m expression.Further, although H2-K appears to be the more highly expressed MHC class I molecule in NOD mice, it would be interesting to know about the expression of H-2D, the 2nd major MHC class I molecule in NOD mice.This also applies to Fig. S4 F and G.
We have also analyzed the protein expression levels of H-2D b in DCs (Fig. 4D and Supplementary Fig. 5E, G).We found that Tyk2 deficiency reduced the expression levels of H-2D b in CD8 + rDC but not other subsets of DCs in the PLN (Fig. 4D (right) and Supplementary Fig. 5E).In the iLN and spleen, reduced expression levels of H-2D b in CD8 + rDC were also observed (Supplementary Fig. 5G).Thus, the expression levels of H-2D b in CD8 + rDC were decreased in Tyk2 -/-NOD mice.
On the basis of these results, we have changed the following text from: "However, we found that Tyk2 deficiency reduced the expression of MHC I (H2-K d ), a crucial molecule for presenting antigens to CD8 T cells 38 , in CD8 + rDC but not other subsets of DCs (Figures 4D and S4F Response: In accordance with the Reviewer's comment, we checked the gene expression data (microarray) of T1D-related autoantigens in β-cells and found that Ins1 and IGRP (G6pc2) were expressed at lower levels in β-cells from Tyk2 -/-mice compared with those from Tyk2 +/+ mice at 6 weeks of age (Supplementary Fig. 1G).At 11 weeks of age, gene expression levels of T1D-related autoantigens including Ins1, Ins2, GAD65 (Gad2), IA-2 (Ptprn), ZnT-8 (Slc30a8), and G6pc2 in β-cells were comparable between genotypes.
On the basis of these results, we have added the following text to the Results: (Line 212-218) "Type I IFN signaling in β-cells was associated with the presentation of autoantigens 19,35 ; therefore, we compared the expression levels of T1D-related autoantigens in β-cells.At 11w, gene expression levels of T1D-related autoantigens including insulin (Ins1 and Ins2), GAD65 (Gad2), IA-2 (Ptprn), ZnT-8 (Slc30a8), and IGRP (G6pc2) in β-cells were comparable between genotypes (Supplementary Fig. 1G).However, at 6w, Ins1 and G6pc2 were expressed at lower levels in β-cells from Tyk2 -/-mice compared with Tyk2 +/+ mice (Supplementary Fig. 1G)." We have also added the following text to the Discussion: (Line 503-509) "In the early phase of T1D development, TYK2 inhibition might reduce the expression levels of islet-associated autoantigens in β-cells (Supplementary Fig. 1G).In contrast to these preservation effects of TYK2 inhibition in β-cells, a recent study suggested that TYK2 inhibition lead to The Tyk2 -/-naïve 8.3 CD8 T cells underwent increased proliferation in the PLN of 24w Tyk2 +/-recipient mice 5d post transfer compared with those in 24w Tyk2 -/-recipient mice (Fig. 4A).Ten days after transfer, the Tyk2 -/-naïve 8.3 CD8 T cells underwent increased proliferation in the PLN of 6 to 8w Tyk2 +/-recipient mice compared with those 5d after transfer (Fig. 4A).
On the basis of these results, we have changed the following text from: "Ten days after the transfer, Tyk2 +/-naïve 8.3 CD8 T cells showed greater proliferation in the PLN of Tyk2 -/-recipient mice compared with 5 days post transfer (Figure 4A).The limited proliferation of 8.3 CD8 T cells was observed in the iLN (Figure S4B), indicating the importance of the PLN for the proliferation of islet-autoreactive CTLs.Together, these observations suggest that CD8 T cell-extrinsic mechanisms have a role in the proliferation of islet-autoreactive CTLs in the PLN." to (Line 264-273) "These observations suggest that CD8 T cell-extrinsic mechanisms have a role in the proliferation of islet-autoreactive CTLs in the PLN.Indeed, Tyk2 -/-8.3CD8 T cells more proliferated in the PLN of Tyk2 +/-recipient mice 10 days after transfer compared with those 5 days post transfer (Fig. 4A).Ten days after transfer, the number of proliferating 8.3 CD8 T cells was increased in the PLN of Tyk2 -/-mice compared with those 5 days after transfer (Fig. 4A), suggesting that the priming of CD8 T cells was impaired but not abolished in the PLN of Tyk2 -/-mice.In agreement with a study showing the importance of the PLN for the proliferation of islet-autoreactive CTLs 39 , the limited proliferation of 8.3 CD8 T cells was observed in the iLN (Supplementary Fig. 4B)."7) Fig. 4B, the authors showed that the proportion of IGRP tetramer positive CD8+ T cells remained stable between 6-wk-old and 11-wk-old mice.Based on the data, the author drew the conclusion that there was no epitope spreading.This reviewer does not think these data support the conclusion, as epitope spreading does not mean the alteration of a tested and disease-causing epitope.
Response: We think that the Reviewer might be mistaken on this point.We described that " (line 275-276) the defective proliferation of CD8 T cells in the PLN of Tyk2 -/-mice is not due to reduced IGRP epitope spreading".
On the basis of the results that the proportion of IGRP tetramer-positive CD8 T cells remained stable between genotypes (Fig. 4B), we considered that IGRP epitope spreading occurred in the Tyk2 -/- NOD mice.If there was no IGRP epitope spreading, IGRP-specific CD8 T cells would not be detected in the PLN because the IGRP epitope is not available.
We have revised the following text in the Results to more clarify our findings.
(Line 274-284) "Because IGRP epitopes were reported to appear and spread in the middle-to-late stage of T1D 42 , defects in IGRP epitope spreading in Tyk2 -/-mice might correlate with the reduced proliferation of 8.3 CD8 T cells in the PLN of Tyk2 -/-mice.To exclude this possibility, we evaluated IGRP-specific CD8 T cells (IGRP CD8 T) using tetramers containing a mimotope of IGRP (NRP-V7).A comparable frequency of IGRP CD8 T cells was observed in the PLN of both Tyk2 genotypes (Fig. 4B).
Furthermore, the proliferation of 8.3 CD8 T cells was also suppressed in the PLN of normoglycemic 24w Tyk2 -/-recipient mice with advanced insulitis compared with 14w Tyk2 -/-mice, but not in the PLN of normoglycemic 24w Tyk2 +/-recipient mice (Fig. 1B and 4A).These data suggest that the defective proliferation of CD8 T cells in the PLN of Tyk2 -/-mice is not due to reduced IGRP epitope spreading in Tyk2 -/-mice, but other factor(s)."8).The authors showed CD44hi T cells, in most if not all, data, and the status of CD62L of the CD44hi T cells was not clear.It is also not clearly stated why the authors focused on CD44hi T cell populations.
Response: In accordance with the Reviewer's comment, we revised texts in the Results to clearly state why we focused on CD44 hi T cells as described below: (Line 255-256) "Antigen encountered naïve T cells become activated, proliferate, and develop into effector and memory T cells that exhibit the CD44 hi phenotype 40 ."(Line 336-338) "Next, we characterized antigen-experienced CD8 T cells that exhibited the CD44 hi phenotype.Because transcription factors (TFs) are key regulators of T cell effector functions, we examined TFs in CD44 hi CD8 T cells in the PLN." We showed the CD44 and CD62L expression status in polyclonal CD8 T cells in the PLN (Fig. 3B and Supplementary Fig. 3C) and pancreas (Supplementary Fig. 3A).However, we did not show these expression status in IGRP-specific CD8 T cells.In the pancreas, CD44 hi IGRP CD8 T cells consisted of a CD62L(-) population only (Supplementary Fig. 6E).In the PLN, CD44 hi IGRP CD8 T cells consisted of CD62L(-) and CD62L(+) populations at a ratio of approximately 6:1 (Supplementary Fig. 6E).
We have added the following text.

Figure 4E :
Figure 4E: the number of labelled DEG confuses the take home message of the figure part, the authors could remove labels of non-ISGs and DEGs irrelevant of DC function

Fig. 2D ,
Fig. 2D, normalized expression levels of B2m and H-2k1 are shown.Because the microarray does not have a probe for H-2d, the gene expression data of H-2d in β-cells are not shown in Fig. 2D.
)." to (line 306-309) "However, we found that Tyk2 deficiency reduced the expression of MHC I (H-2K d and H-2D b ), a crucial molecule for presenting antigens to CD8 T cells 40 , in CD8 + rDC but not other subsets of DCs (Fig. 4D and Supplementary Fig. 5E, G)." Revised Fig. 4D.Expression levels of H-2K d and H-2D b in CD8 + rDC in the PLN.P-values were calculated using Kruskal-Wallis test with Dunn's posttest.Revised Supplementary Fig. 5E.Expression levels of H-2D b in DCs in the PLN.P-values were calculated using Kruskal-Wallis test with Dunn's posttest.Revised Supplementary Fig. 5G.Expression levels of H-2D b in CD8 + rDC in the iLN and spleen.P-values were calculated using Kruskal-Wallis test with Dunn's posttest.5).In the context of the lower expression of Fas and H2-K, what happened to beta cell function, such as insulin production?Was this also changed in the absence of Tyk2?This question could shed light on whether beta cells are more self-protected by down regulating molecules related to autoimmune attack or the beta cells may also down regulate production of self antigen(s).