DOK1 and DOK2 regulate CD8 T cell signaling and memory formation without affecting tumor cell killing

Targeting intracellular inhibiting proteins has been revealed to be a promising strategy to improve CD8+ T cell anti-tumor efficacy. Here, we are focusing on intracellular inhibiting proteins specific to TCR signaling: DOK1 and DOK2 expressed in T cells. We hypothesized that depletion of intracellular inhibition checkpoint DOK1 and DOK2 could improve CD8+ T-cell based cancer therapies. To evaluate the role of DOK1 and DOK2 depletion in physiology and effector function of CD8+ T lymphocytes and in cancer progression, we established a transgenic T cell receptor mouse model specific to melanoma antigen hgp100 (pmel-1 TCR Tg) in WT and Dok1/Dok2 DKO (double KO) mice. We showed that both DOK1 and DOK2 depletion in CD8+ T cells after an in vitro pre-stimulation induced a higher percentage of effector memory T cells as well as an up regulation of TCR signaling cascade- induced by CD3 mAbs, including the increased levels of pAKT and pERK, two major phosphoproteins involved in T cell functions. Interestingly, this improved TCR signaling was not observed in naïve CD8+ T cells. Despite this enhanced TCR signaling essentially shown upon stimulation via CD3 mAbs, pre-stimulated Dok1/Dok2 DKO CD8+ T cells did not show any increase in their activation or cytotoxic capacities against melanoma cell line expressing hgp100 in vitro. Altogether we demonstrate here a novel aspect of the negative regulation by DOK1 and DOK2 proteins in CD8+ T cells. Indeed, our results allow us to conclude that DOK1 and DOK2 have an inhibitory role following long term T cell stimulations.


Mice and cell lines
Mice Generation of Dok1/Dok2 DKO mice was previously reported 33 .Pmel-1 TCR transgenic mice, were a kind gift from Nicholas P. Restifo (NCI, Bethesda, USA) 34 .Mice were crossed to generate Dok1/Dok2 DKO Pmel-1 transgenic mouse strain.C57BL/6 Ly5.1 were purchased from Janvier Labs and housed in our animal facility at least one week before starting the experimental protocol.All mice were crossed, housed and genotyped according to the guidelines of Committee for Animal Experimentation of Marseille and in accordance with European Directive 2010/63/EU.The experimental protocol was approved by an Institutional Animal Care and Use Committee (the Structure du bien-être animal (SBEA) du Centre de Recherche en Cancérologie de Marseille (CRCM) and the Comité d'Ethique en Expérimentation Animale n°14 - CEEA 14).Male and female mice were used between the ages of 6-12 weeks.Mice were bred and maintained under specific pathogen-free conditions at the Centre de Recherche en Cancérologie de Marseille (CRCM) animal facility.

Cytotoxicity
Target B16 cells were stained with 4 μM of Cell Proliferation Dye eFluor™ 670 (Life Technologies) according to manufacturer's instructions.Primed CD8 + T cells were then incubated with target cells for four hours at 37 ℃ at different effector to target (E:T) ratios.Target cell killing was measured using CellEvent™ Caspase-3/7 Green Detection Reagent (Life Technologies) and analyzed by flow cytometry.

Conjugate formation
This method has been adapted from a previous report on NK cells 36 .Here, primed CD8 + T cells were incubated for 30 min on ice with CD8-APC-EF780 antibody (Invitrogen, #47-0081-82) in serum-free RPMI medium.They were then washed and resuspended at 20.10 6 cells per ml. 100 μL of cell suspension was then added to 100 μL of labeled with Cell Trace Violet (V450) (Invitrogen) B16 WT or B16-hgp100 cells (at 20.10 6 cells per mL) and centrifuged at 1,500 rpm (4 ℃).After removing 150 μL of supernatant, cells were stimulated by incubation at 37 ℃ for 0, 5, or 10 min.Reactions were stopped by adding ice-cold phosphate-buffered saline.Conjugates were detected by flow cytometry as double positive CD8 + V450 + events.

In vivo migration assays
Primed CD8 + T cells isolated from WT or Dok1/Dok2 DKO mice were loaded with Cell Trace Violet Stain (Life Technologies #C34557) or Cell Trace Far Red DDAO (Life Technologies #C34553).As described previously 37 , cells were mixed at a ratio of 1:1 and 10.10 6 cells were i.v.injected in C57BL/6 mice.Then, 1 h later, recipient mice were euthanized, and blood, spleen, and (inguinal, axillary and brachial) lymph nodes were removed for quantification of Cell Trace Violet-labeled and Cell Trace Far Red-labeled T cells by flow cytometry.

Statements
The study is reported in accordance with ARRIVE guidelines ( https:// arriv eguid elines.org ).Furthermore, the experiments were conducted in accordance with the Guidelines of the European Union Council (2010/63/EU) and French legislation for the use of laboratory animals (project authorization of animal experimentation: DAP #28902).

Primed Dok1/Dok2 DKO CD8 + T lymphocytes have an effector/memory phenotype
To understand how DOK1 and DOK2 regulate CD8 + T cell activity and especially their cytotoxic function against cancer cells, we crossed Dok1/Dok2 DKO mice with pmel-1 TCR transgenic mice.We first tested the impact of DOK1 and DOK2 deletion on naïve and in vitro amplified CD8 + T cells.Dok1/Dok2 DKO and WT resting CD8 + T cells show similar proportions of CD4 + and CD8 + T cells, naïve (CD62L + CD44-), central memory (CD62L + CD44 +) and effector memory (CD62L-CD44 +) T cell subsets.To prime cells, naïve CD8 + T lymphocytes from spleen were purified and then expanded for 5 days, with anti-CD3, anti-CD28 or hgp100 and IL-2 for 3 days followed by 2 days in IL-2 only.Similar protocol using splenocytes is applied for hgp100 peptide treatment (Fig. 1A).T cell subset phenotype was followed over the time at day 3 and day 5 by flow cytometry.Although no difference in proportion of Naïve (CD44-CD62L +), Central memory (CD44 + CD62L +) and Effector/Memory (CD44 + CD62L-) was observed in unstimulated CD8 + T cells or at day 3 of expansion, (Fig. 1B) we noticed that Dok1/Dok2 DKO CD8 + T cells had a higher proportion of effector memory cells compared to WT cells at the day 5 of expansion (Fig. 1C).We identified the difference of CD62L expression in WT and Dok1/Dok2 DKO CD8 + T cells between day 3 and day 5 (Fig. 1D).Therefore, DOK1 and DOK2 regulate the formation of memory CD8 + T cells.

Dok1 and Dok2 invalidation improves TCR signaling in primed CD8 + T cells
We, then, sought to explore the role of DOK1 and DOK2 invalidation in TCR signaling.We used two doses (5 and 1 μg/mL) of biotinylated anti-CD3 to determine the optimal dose for CD8 + T cell stimulation (Fig. S1C).Dok1/Dok2 DKO and WT CD8 + T cells purified from spleen were stimulated with biotinylated anti-CD3 and cross-linked with streptavidin during the indicated time.After cell lysis the levels of pErk and pAkt were evaluated by Western blot (WB) analysis.Naïve WT and DKO CD8 + T cells show the same level of pErk and pAkt upon TCR stimulation (Fig. 2A).
To determine whether the loss of DOK1 and DOK2 affects TCR signaling in primed CD8 + T cells, similar experiments were performed.Primed Dok1/Dok2 DKO CD8 + lymphocytes showed an upregulation of pErk and pAkt expression compared to WT CD8 + T cells upon TCR stimulation, although only pAkt appeared to be statistically significant (Fig. 2B).Subsequently, phosphoflow experiments were performed.Primed CD8 + T cells were stimulated with biotinylated anti-CD3 and cross-linked with streptavidin during 5 min.Phosphorylation of ERK1/2 was detected by flow cytometry (Fig. 2C).Again, pErk expression was increased in primed Dok1/Dok2 DKO CD8 + T cells compared to WT cells, confirming our WB experiments.
To ensure that this TCR signaling improvement in Dok1/Dok2 DKO primed CD8 + T lymphocytes is conserved with a different TCR stimulation setting, we performed a stimulation with hgp100 peptide recognized by the Pmel-1 TCR.Cells were peptide-stimulated, immediately fixed and stained with antibodies against pErk and pS6.Flow cytometry analysis revealed a small increase of pErk induction in Dok1/Dok2 KO primed CD8 + T lymphocytes compared to WT CD8 + T cells (Fig. 2D).No difference in pS6 expression (Fig. 2E) or in CD69 cell surface expression (Fig. S1B) was found.
Similarly, we performed immunoblot analysis of primed CD8 + T cells lysates after a stimulation with hgp100 peptide.We did not notice any difference in pErk and pAkt expression levels between WT and DKO cells (Fig. S2A-C).
Altogether, these findings suggest that DOK1 and DOK2 deficiency enhances TCR signaling especially upon CD3 mAb stimulation.This effect was only observed when CD8 + T cells were primed in vitro.

Dok1/Dok2 DKO and WT CD8 + T cells show similar cytotoxicity in vitro
To assess CD8 + T cell cytotoxicity a murine B16 melanoma cell line expressing constitutively hgp100 antigen was used.Primed pmel-1 + CD8 + T cells (corresponding here to the primed WT CD8 + T cells) can recognize hgp-100 antigen at the surface of B16-hgp-100 expressing cells but not when the peptide is not expressed.Primed Dok1/Dok2 DKO and WT CD8 + T cells were co-cultured with B16-hgp100 cells at indicated effector/Target (E/T) ratio.Expression of IFN-γ and TNF-α was detected by flow cytometry (Fig. S3A).Surprisingly, Dok1/Dok2 DKO and WT CD8 + T cells expressed the same level of IFN-γ and TNF-α for all tested E/T ratios (Fig. 3A  not shown).Likewise, degranulation marker CD107a showed also similar expression between Dok1/Dok2 DKO and WT CD8 + T cells (Fig. 3B).
Next, the capacity of primed CD8 + T cells to kill B16-hgp100 cells was analyzed by measuring caspase 3/7 activation in target cells after 4 h of co-culture.Whatever the ratio tested, WT and Dok1/Dok2 DKO primed CD8 + T cells displayed a similar cytotoxic activity (Fig. 3C).
Therefore, these data suggest that Dok1/Dok2 DKO and WT CD8 + T cells show similar cytotoxicity in vitro.

cells do not improve survival of tumor-bearing
Finally, we evaluated the transfer of primed CD8 + T cells to provide a significant benefit survival for tumorbearing mice.We made a subcutaneous B16-hgp100 tumor injection and 10 days after we performed an adoptive cell transfer of WT and Dok1/Dok2 DKO CD8 + T cells.PBS injection was used in a control group.We found that adoptive cell transfer improved a survival of tumor bearing mice, however, there was no difference between WT and Dok1/Dok2 DKO group (Fig. S4).In line with the in vitro experiments, Dok1 and Dok2 invalidations do not improve primed CD8 + T cell cytotoxicity in the context of adoptive cell transfer.

Dok1/Dok2 DKO and WT CD8 + T cells show similar effector: target conjugate formation and migration.
Since, Dok1/Dok2 DKO primed CD8 + T lymphocytes have fewer expression of CD62L adhesion molecule (Fig. 1D), we hypothesized that it could impact the formation of effector-target cell conjugates or migration.For migration experiment WT and Dok1/Dok2 DKO primed CD8 + T cells were labeled with dyes of different color.Next cells were mixed in proportion 1:1 and 10 6 of cells were intravenously injected in healthy mice.One hour after injection blood, spleen and lymph nodes were taken to evaluate the proportion of WT and DOK-1/2 DKO CD8 + T in each organ.No difference in cell proportion was detected (Fig. S5 A-D).To evaluate conjugate formation, we stained target cells with Cell Trace Violet and effector cells with fluorochrome-coupled CD8 mAb.A co-culture experiment was performed, and conjugate formation was followed during time course (Fig. S6A-C).WT and Dok1/Dok2 DKO primed CD8 + T cells showed the same effector-target conjugate formation.Similar induction of pAkt and pErk are detected by immunoblots at early time points (Fig. 2B & S2B,C) by using a peptide stimulation, however we were unable to detect a boost of TCR signaling due to the absence of DOK proteins.
In primary T cells, TCR engagement with a CD3 antibody is orders of magnitude stronger than cognate recognition of peptide-MHC 46 .In a T cell line, it has been suggested that after a strong TCR stimulation, possible mechanisms of negative feedback regulation could dampen TCR signaling 47 .This could explain why the upregulation effect of DOK deficiency is only visible upon antibody stimulation.We can notice that a slight pErk upregulation can be detected by phosphoflow analysis (Fig. 2D) upon peptide stimulation in DKO condition.This could be due to the sensitivity experimental method as for flow cytometry, the phosphorylation events are detected at the single cell level 48 .
Naïve and memory T cells have considerable differences in their function physiology and TCR signalosome.Particularly, it was demonstrated that memory CD8 + T cells have more CD8-bound Lck than naïve cells and CD4 + T cells have less Zap70 and Slp76 phosphorylation upon TCR stimulation, suggesting a faster and more efficient signal transduction pathway in memory T cells 17,19 .These data confirm our findings and suggest some functional or structural differences in CD8 + naïve versus primed signalosomes.Further investigation notably using high throughput technologies such as mass spectrometry is needed to decipher these phenomenona 11 .
Compensation mechanisms and signaling re-wiring may also occur in TCR signaling pathway, like the inhibitory TCR signaling protein Csk, normally associated with PAG, could associate with another protein PTPN22 in absence of PAG compensating TCR signaling 40 .DOK1 and DOK2 could be seen as a platform to recruit other inhibitory proteins (RasGAP, SHIP, Csk).Maybe other proteins could compensate the lack of DOK1 and DOK2 when these proteins are totally absent.Thus, methods to downregulate transiently DOK1 and DOK2 in CD8 + T lymphocytes by shRNA or CRISPRi-Cas9 techniques may be interesting to avoid these compensation mechanisms and understand more precisely DOK1 and DOK2 regulation.
In this study, we wanted to assess the capacity of TCR signaling inhibitory proteins DOK1 and DOK2 to improve CD8 + T cells immunotherapy.In the development of T cell-based immunotherapy the problem of T cell functionality blunting in the tumor microenvironment is crucial.The concept of improving the strength of TCR signaling upon TCR activation is very important to overcome this problem.Therefore, we tested TCR signal inhibiting proteins DOK1 and DOK2 as potential candidates to increase TCR signaling in CD8 + T cells.The fact that we found the increase of pErk and pAkt in primed but not naïve CD8 + T cells is even advantageous in this context, as only primed CD8 + T cells are used for adoptive cell transfer immunotherapies nowadays.Previously, it was shown that inhibition of Akt pathway by rapamycin could improve the generation of memory cells in terms of their quantity and quality 49 .The acquisition of effector functions of CD8 + T cells associated with intense Akt signaling impairs the in vivo antitumor efficacy of adoptively transferred cells 50 .The emerging consensus on this question is that central memory tumor-reactive CD8 + T cells have an improved antitumor capacity in comparison with effector memory cells 51,52 .Therefore, the activation both pErk and pAkt as we can see in the context of Dok1/Dok2 invalidation would be not advantageous for antitumor capacity of primed CD8 + T cells as the positive influence of pErk upregulation would be compensated by the negative influence of effector memory phenotype due to increased pAkt.Probably in the concept of CD8 + T cells immunotherapy improvement by acting through TCR signaling inhibiting such polyvalent inhibiting proteins as DOK1 and DOK2 would be excessive, and proteins acting on inhibiting of one specific signaling pathway would fit more.As successful examples of targeting intracellular inhibiting proteins in the context of cancer immunotherapy, we can mention recently adapted for clinical trials CISH, targeting PLC-γ1 and HPK-1, targeting SLP-76 9,22,53 .Both could be associated to Erk pathway improvement, without direct effect on Akt pathway.The role of DOK1 and DOK2 has been reported in another type of cytotoxic lymphocytes, the Natural Killer (NK) cells 54 .It would be interesting to challenge the role of these DOK adaptors in NK cells to eliminate cancer cells.
In summary, our data provided evidence that DOK1 and DOK2 interfere in primed CD8 + T cell TCR signaling negative regulation and have impact on memory CD8 + T cell formation.We underlined an interesting phenomenon that DOK1 and DOK2 could play a different role in naïve and memory TCR signaling, however based on our model the DOK1/DOK2 adaptor proteins do not appear to be good candidates for CD8 + T cell manipulation in immuno-oncology.Therefore, due to complexity of TCR signaling there is a real need of screening studies of invalidation of TCR signaling inhibitory proteins to improve existing CD8 + T cell-based immunotherapies.
Dok1/Dok2 DKO or WT Ly5.2 primed CD8 + T cells and received intraperitoneal injections of human IL-2 (Aldesleukin, Clinigen, NL) in PBS (6.10 4 IU/0.5 mL) once daily for 3 days starting on the day of T cell transfer.As a control, IL-2 is also injected in the "PBS" conditions.Mice with tumors greater than 400 mm 2 or in illness state were euthanized (Application for a project authorization of animal experimentation: DAP #28902).