Field of the invention

The present invention relates to a method for the rapid in vitro generation of viable, exhausted T cells and related cell cultures and applications of said method for drug discovery.

Background of the invention

Exhausted T cells are effector T cells with decreased cytokine expression and effector function. Exhausted T cells are deemed to be resistant to reactivation and exhibit decreased proliferation. T cell exhaustion occurs naturally when the cells are chronically activated at sites of chronic inflammation, such as cancer, autoimmunity, and chronic infection. In a review by Crespo et al., T cell exhaustion was compared to T cell anergy, senescence and sternness in a tumor environment (Crespo et al., Curr Opin Immunol. 25(2): 214-221, 2013). Several markers and features have been identified that are required to distinguish T cell exhaustion from other T cell dysfunctions. Exhausted T cells show a significant decrease in the cytokines IL-2, IFN _Y , and TNFa expression (E. John Wherry et al., Immunity 27(4), 670-684, 2007), as well as cell cycle arrest. A hallmark of exhaustion is the loss of poly-functionality which means that an individual cell looses the capacity to make multiple cytokines simultaneously (E. John Wherry et al, Immunity 27(4), 670-684., 2007). Another critical feature of exhausted CD8 T cells is the sustained high expression of multiple inhibitory receptors, including PD-1, Tim-3a, Lag-3, CD160, CD244,TIGIT and others (Jin, H. T. et al. Proc. Natl Acad. Sci. USA, 107 (33): 14733-14738,2010; Johnston, R. J. et al, Cancer Cell, 26 (6): 923-937, 2014 ). Characteristic of exhausted CD8 ⁺ T cells is also their deficient ability to proliferate. At the molecular level altered expression of transcription factors is a defining feature of exhausted CD8 ⁺ T cells (Kao, C. et al. Nat. Immunol. 12(7), 663- 671,2011), for example changes in expression of Tox and TCF1 can discriminate between fully exhausted T cells and those that retain proliferative capacity (B, C. Miller et al., Nat. Immunol. .20 (3), 326-336, 2019). It has been recognized that T cell exhaustion has an extremely important role in both cancer and chronic infections. However, no adequate in vitro model is yet available, which would allow researchers to investigate exhausted T cells without any other interfering. A recent attempt has been made by D. Ozkazanc et al.

(Immunology, 149, 460-471, 2016) alleges that T cells co-cultured with leukemia cells under continuos stimulation become exhausted. However, the majority of the cells (about 80%) are not exhausted but anergic. Only about 20% of cells were exhausted with this method. As described above, exhaustion is characterized by co- expression of multiple inhibitory receptors. In Ozkazanc et al. co-expression of Tim3 and LAG3 were observed in 22.8% or 18.7% of cells. Importantly, the authors show that IL-2 restores proliferation of these cells, wich is a characteristic of anergy and not of exhaustion. It is well known in the art that IL-2 prevents anergy (Beverly B et al, Int. Immunol. 4:661,1992; Macian F et al, Cell, 109: 719, 2002; Schietinger A et al, Trends Immunol. 35: 51, 2013). Further, the cells described by Ozkazanc et al. are still capable of making cytokines TNFa and IL-2 and only show a 50% reduction, which also does not support exhaustion. Finally, Ozkazanc et al. does not provide any evidence that a defining characteristic of exhaustion, i.e. the expression of the transcription factor Tox, is expressed in the cell cultures.Another attempt was made by the group of Junghans, who identified YY1 as a down regulator of type I cytokines and an upregulator of checkpoint receptors in exhausted T cells (M.Y. Balkhi et al., iScience 2, 105-122, 2018). The researchers showed that their in vitro model successfully has decreased IL-2 and IFN _g production, and that PD1, Tim3 and Lag3 have an increased expression. However, neither loss of polyfunctionality has been shown, nor the deficient capacity to proliferate in vivo. Notably, IL-2 production was reduced but still present while the IL-2 promoter was highly active making IL-2, indicating that these cells are not exhausted. Further, the weak IFN _g production in these cells also does not support T cell exhaustion. Further, only 68% of cells express PD-1, the cardinal receptor for exhausted T cells, which argues for at best partial exhaustion. There is no evidence for co-expression of inhibitory receptors and no data on loss of polyfunctionality loss. The induction of YY1 in these cultures suggests that differentiation of T cells is being affected and not exhaustion. YY1 regulates Th2 cell differentiation via the master regulator GATA3 and the Treg inducer Foxp3 so that it is likely instead of exhaustion the method of Balkhi et al. induces Th2 or Treg cells instead of exhaustion. Hence, because these described changes are compatible with differentiation of naive T cells into effector T cells, the loss of cytokine production and especially IL-2 which is made by naive T cells but not effector T cells is an insufficient criterion to define T cell exhaustion. The above study thus also failed to demonstrate the important changes in the expression of transcription factors that characterize exhaustion. Additionally, the in vitro generated T cells are not described as viable, i.e. preventing them from being cleared from the medium, thus reducing their applicability in in vitro experiments.

Accordingly, there is still need for a method for generating viable, exhausted T cells.

Summary of the invention

It is an aim of the present invention to provide a method to rapidly generate viable, exhausted T cells in vitro. The present inventors have found that this can be achieved through repeated antigen stimulation of T cells in the presence of IL-7 and/or IL-15. More particularly, the invention comprises a method for in vitro generation of large numbers of viable, exhausted T cells, comprising culturing T cells, especially CD8 ⁺ T cells, in a medium comprising at least IL-7 and/or IL-15; and antigen stimulation of said T cells to generate exhausted T cells. The invention provides methods and means to induce exhausted T cells rapidly over a few days, for example 5 days, a significant improvement for intervention testing as exhaustion in vivo can take 30 days or more to be fully induced.

The approach taken allows to establish exhaustion as a result of T cell receptor stimulation without the presence of other factors that can influence gene expression or behavior of these T cells. It allows in vitro screening of compounds, drugs or other approaches to induce, prevent, accelerate or reverse T cell exhaustion.

Further provided herein is a method for testing treatment that prevents, alleviates or accelerates exhaustion of T cells, comprising culturing T cells in a medium comprising at least IL-7 and/or IL-15; and antigen stimulation of said T cells to generate exhausted T cells; wherein said T cells are treated before, during and/or after the culturing and/or the stimulation. In one embodiment, treatment can include administration of a chemical compound, such as a small molecule, a peptide, a protein (including antibodies), a nucleotide and/or a carbohydrate.

Furthermore, or in addition, treatment can comprise genetic modification, including gene editing, of the T cells.

Said antigen preferably is a peptide, more preferably a specific peptide is selected from the group of globular proteins and fragments thereof, preferably ovalbumin and fragments thereof, most preferably OVA257-264 peptide.

A total concentration of IL-7 and/or IL-15 together is 1- 100 ng/mL, preferably 5-50 ng/mL, most preferably 10-25 ng/mL.

Said antigen concentration preferably is 1 ng/mL to 10 mg/mL, preferably 5 ng/mL to 5 mg/mL, most preferably 10 ng/mL to 1 mg/mL.

Said T cells preferably are cultured for 2-15 days, more preferably for 3-10 days, most preferably for about 5 days.

The generated T cells preferably are evaluated in vivo, comprising the steps of in vitro generation of exhausted T cells according to the methods of the invention; transfer of the exhausted T cells into an animal; stimulation of the animal with the antigen; and detection of exhaustion markers.

The aforementioned animal preferably is a human or a rodent, most preferably a mouse.

Further provided herein is a method for identifying molecular targets to prevent, restore, or accelerate T cell exhaustion comprising culturing T cells in a medium comprising at least IL-7 and/or IL-15; repeated antigen stimulation of said T cells to generate exhausted T cells; treating said T cells with a treatment that may prevent, restore, or accelerate T cell exhaustion; and comparing the treated T cells with untreated exhausted and non-exhausted T cells, to find relevant changes in said T cells.

Further provided herein is a cell culture comprising viable exhausted T cells as obtained by a method according to the invention.

The provision of a method to generate viable exhausted T cells in vitro, is a less time-consuming method than in vivo methods and yields more experimental material. Additionally, the environment of these in vivo models such as

inflammation, high viral loads, suppressive cytokines or suppressive cells can all obscure the phenotype of exhausted T cells making it hard to dissect the true exhaustion-associated changes. A possible solution to these problems is the establishment of an in vitro model to induce exhausted T cells. Establishing an in vitro model can not only speed up the process of driving the cells to exhaustion but also produce abundant experimental materials with fewer animals. It will also provide a tool to manipulate genes and screen the potential therapies to restore the function of the exhausted CD8 ⁺ T cells. The development of an in vitro model therefore could greatly facilitate the study of T cell exhaustion and allows for the testing of novel approaches to reverse exhaustion.

The above and other characteristics, features and advantages of the concepts described herein will become apparent from the following detailed description, which illustrates, by way of example, the principles of the invention.

Legends to the figures

Figure 1 . Scheme of the experimental protocol and numbers of live exhausted CTL generated by in vitro exhaustion. A) Scheme of the experimental set up. B) Purified CD8+ T cells (0.5x10 ⁶ cells per well) were cultured either unstimulated, stimulated once with peptide and repeat peptide stimulated.. Live cells were counted on day 5. Pooled data showing absolute cell numbers (left) and fold expansion (right). Data are from n=13 and 10 independent experiments. C) Purified CD8+ T cells (0.5x10 ⁶ cells per well) were cultured in presence of IL-7 alone, IL-15 alone or a combination of IL7 and IL-15 and either unstimulated, stimulated once with peptide or repeat peptide stimulated. Live cells were counted on day 5.

Figure 2. In vitro exhaustion method induces loss of cytokine production and polyfuctionality. Purified OT-I CD8+ T cells were cultured either without peptide (no peptide), stimulated one time for 2 days with OVA peptide (single peptide) or stimulated with OVA peptide daily (repeat peptide). On day 5 cells were harvested and re-stimulated with OVA peptide. A) Pooled data showing the frequency of cytokine producing cells after re-stimulating with OVA peptide. B) SPICE figures depicting the frequency of cells producing one, two or three cytokines in different combinations. Each symbol represents one animal (n=ll-12), 9 independent experiments performed. Line depicts mean ± SE.

Between the groups, Student’s t test with Welch’s correction was performed except for % IFN- _U + (ANOVA with Tukey’s post hoc test). *<0.05, **P<0.01, ***P<0.001,

****p<0.00001.

Figure 3. Repeat peptide stimulated CD8+ T cells have decreased cytotoxic function. On day 5, cells were harvested and re-stimulated with OVA peptide. (A) Median fluorescence intensity (MFI) of the degranulation marker, CD107a, shown (left panel). Fold change of CD107a MFI induced by peptide re- stimulation depicted on the right panel. (B) MFI of Granzyme B (GzmB) depicted for the different culture conditions. Each symbol represents one animal (n=ll-12),

9 independent experiments performed. (C) No peptide, single peptide and repeat peptide stimulated cells were co-culture with target cells (OVA-pulsed AE-17 cells) at different ratios. Percentage of specific killing is depicted. One of five

independent experiments shown. Line depicts mean ± SE. Between the groups, Student’s t test with Welch’s correction was performed except for CD 107a MFI (Wilcoxon signed rank test). *<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Figure 4. In vitro exhaustion method induces the expression of inhibitory receptors. A. CD8 ⁺ T cells (CTL) submitted to in vitro exhaustion (multiple peptide) express multiple inhibitory receptors. Data are for inhibitory receptors PD-1, Lag3, Tim3a, CD 160, TIGIT and CD244. B. SPICE figures depicting the frequency of CD8+ T cells expressing one, two, three or four inhibitor receptors (PD-1, CD244, CD160, Lag3) simultaneously. Each symbol representative one animal (n=7-ll), with 7-9 independent experiments performed. Line depicts mean ± SE. Between the groups, Wilcoxon signed rank test was performed with exception of % Tim-3+ and % CD 160+ for which a Student’s t test with Welch’s correction was used. *<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Figure 5. In vitro exhaustion method induce a distinct Transcription Factor expression pattern. Expression level of transcription factors Tcfl, Tox, EOMES and T-bet in live CD8+ T cells after five days of different treatments.

Every symbol corresponds to one animal. Error bars show SEM. ***P<0.001, ****P<0.0001.

Figure 6.In vitro exhaustion method induces CTL with reduced capacity to proliferate in vivo. OT-I CD8 ⁺ T cells (CTL) submitted to in vitro exhaustion (multiple peptide) were transferred into mice and then mice were intranasally infected with an influenza virus that expresses the OVA peptide recognized by the OT-I cells. None stimulated or CTL stimulated only once expand in the mouse and large numbers are found in the lungs of mice. In vitro exhausted (multiple peptide) expand much less indicating a defect in proliferative capacity. a and b show percentage and absolute numbers of donor cells in lung. c and d show the endogenous response of the host mouse, showing that in vitro exhausted (multiple peptide) cannot compete as well as the other cells and hence the endogenous response is higher. Data from n=8 mice and 3 experiments.

Figure 7. In vitro exhaustion method induces CTL whose

transcriptome is enriched for gene sets associated with exhaustion. A) Principle component analysis of RNAseq of no peptide , single peptide and multiple peptide stimulated cells reveals that CD8 ⁺ T cells (CTL) submitted to in vitro exhaustion (multiple peptide) cluster separately from the other cells. B) Analysis of RNAseq data from OT-I CD8 ⁺ T cells (CTL) submitted to in vitro exhaustion (multiple peptide) shows that these cells have significantly enriched for genes that are associated with CTL exhaustion. C) Igenuity pathway analysis (IPA) of gene expression shows that in vitro exhaustion (multiple peptide) the signaling pathway that is most upregulated is the T cell exhaustion pathway. RNAseq data from n=4- 5 mice and 4-5 experiments.

Figure 8; Genome wide DNA methylation changes during T cell stimulation reveal Tcf7 promotor methylation. Sorted live CD8+ T cells were processed and whole genome methylated DNA sequencing (MeD-seq) was performed. (A) Hierarchical clustering on DMRs (differentially methylated regions) found between the three different peptide exposure conditions are shown. (B) Boxplot of DNA methylation read count data in a 2kb window surrounding the TSS of mentioned genes are shown. The samples were collected from three independent experiments. (C) Representative histogram of TCF1 expression (left) and pooled data showing TCF1 MFI on live CD8 cells (right) in the presence or absence of 20mM DNMT inhibitor during the last 3 days of repeat peptide stimulated cells shown. Data are from n=6 performed in 3 independent experiments, paired

Student’s t test performed. **P<0.01. Figure 9: DNMT inhibitor decreases expression of inhibitor receptors on the repeat peptide stimulated cells. Pooled data showing day 5 inhibitor receptor MFI on repeat stimulated cells in the presence or absence of 20mM DNMT inhibitor. Inhibitor was added during the last 3 days of culture. Data are from n=6 animals, performed in 3 independent experiments. Line depicts mean ± SE. To determine significant differences between the different treatment, paired t-test was used *P<0.05, ** P< 01, ***P<0.001.

Figure 10: Exposure to IL-2 does not alter the exhaustion phenotypes in repeat peptide stimulated cells and CD80 is upregulated on repeat stimulated cells. CD8+ T cell exhaustion was induced in the presence of IL-2 (20U/ml) and IL-7/IL-15 (5ng/ml each). Cytokine production after OVA peptide restimulation for 6 hours (A) and inhibitory receptor expression on day 5 (B) are shown. n=8 animals from 5-6 independent experiments depicted. Representative histogram of CD80 expression on the cells is shown in (C). One of two independent experiments shown (n=3).

Figure 11: Anergy gene sets and pathways do not characterize repeat peptide exhausted T cells. Both upregulated and downregulated genes in anergic T cells are enriched in exhausted T cells. A. GSEA of genes upregulated and B. downregulated in anergic T cells (gene set GSE 5960) are both enriched in differentially expressed repeat peptide stimulated cells (repeat peptide versus no peptide cell). C. GSEA of genes upregulated and D. downregulated in anergic T cells (gene set GSE 5960) are both enriched in differentially expressed Day 30 gp33-specific CTL from LCMV clone 13 infected animals (Day 30 versus gp33- specific CTL naive CD8+ T cells; data from GSE41867). E. In repeat peptide stimulated cells the anergy related pathway in IPA analysis was neither upregulated not downregulated. IPA analysis was performed on the differentially expressed genes in repeat stimulated cells (compared to no peptide). The “Regulation of IL-2 expression in activated and anergic T lymphocytes” pathway is showed enlarged and the overall positioning in the global IPA pathway analysis is indicated by the arrow and red circle. Significantly upregulated pathways (orange) and downregulated pathways (blue), no activity pattern available (grey) are depicted. Bar graphs depict P values (hypergeometric test) presented as -Log10. Detailed description of the invention

While potentially serving as a guide for understanding, any reference signs used herein and in the claims shall not be construed as limiting the scope thereof.

As used herein, the singular forms“a”,“an”, and“the” include both singular and plural referents unless the context clearly dictates otherwise.

The terms“comprising”,“comprises” and“comprised of’ as used herein are synonymous with“including”,“includes” or“containing”,“contains”, and are inclusive or open-ended and do not exclude additional, non-recited members, elements or method steps. The terms“comprising”,“comprises” and“comprised of’ when referring to recited components, elements or method steps also include embodiments which“consist of’ said recited components, elements or method steps.

Unless otherwise defined, all terms used in disclosing the concepts described herein, including technical and scientific terms, have the meaning as commonly understood by one of ordinary skill in the art. By means of further guidance, definitions for the terms used in the description are included to better appreciate the teaching of the present disclosure. The terms or definitions used herein are provided solely to aid in the understanding of the teachings provided herein.

The term“T cell” as used herein refers to lymphocytes developed in the thymus, and can be distinguished from other lymphocytes by the presence of a T- cell receptor on their surface. Although validation of the current described methods have been performed solely on CD8 ⁺ T cells, the current invention should be in no way be considered limited to CD8 ⁺ T cells, but also comprises, but is not limited to, CD4 ⁺ T cells, memory T cells, non-differentiated T cells and the like.

The term“exhaustion” as used herein refers to the dysfunctional state of T cells when the expression of cytokines and the effector function are decreased, and the T cells are resistant to reactivation and exhibit reduced proliferative capacity.

It has been shown that a significant decrease in IL-2, IFN _g and TNFa expression as well as cell cycle arrest are the hallmarks of T cell exhaustion. In tandem with this, an increased expression of inhibitory receptors is also observed, such as the receptors PD-1, Lag3, CD160, CD244, Tim-3a and TIGIT. The term“cell culture”, as used herein, refers to an in vitro population of viable cells under cell cultivation conditions, i.e. under conditions wherein the cells are suspended in a culture medium that will allow their survival and preferably their growth. The cell culture is usually comprised in a container holding the cell culture, referred to as a culture chamber, wherein sufficient exchange of gases such as oxygen and CO ₂ is allowed between the cell culture and the atmosphere to support cell viability.

The term“medium”, as used herein, refers to a medium generally used in the culturing of mammalian T cells. Any available medium suitable for the culturing of mammalian T cells can be utilized, including a medium based on Dulbecco's Modified Eagle Medium (DMEM) or Ham’s F12 nutrient medium, or combinations of such basal media, optionally supplemented with compounds that prevent bacterial contamination, such as antibiotics, preferably in the form of a

combination of penicillin/streptomycin/glutamine. Preferably, the medium comprises Roswell Park Memorial Institute (RPMI) medium (commercially available from e.g. ThermoFisher Scientific, Paisley, UK).

The term“antigen stimulation” as used herein, refers to the exposure of T cell receptors on the surface of T cells to their respective antigen or antigen fragments. The term“repeated antigen stimulation” as used herein, refers to antigen stimulation during multiple days, preferably at least 5 days, more preferably 5 to 10 days, more preferably 5 to 8 days, more preferably 5 to 7 days. Antigen are for instance added to the culture medium once a day for the indicated number of days. Antigens or antigen fragments, should here be understood as molecules

recognizable by T cell receptors which are able to bind to the T cell receptor and stimulate the T cell receptor. During the repeatsed antigen stimulation the cells are culture in a medium comprising at least IL-7 and/or IL-15.

The terms“IL-7” and IL-15”, as used herein, refer to two cytokines. IL-7, or interleukin- 7 is a growth factor for both T- and B-cells in an early lymphoid stage. IL-15, or interleukin- 15 comprises multiple functions, including the proliferation of T cells. IL-7 and IL-15 can be obtained from any source, preferably from mice. Most preferably, murine IL-7 and IL-15 can be commercially obtained from Peprotech (London, UK), catalogue numbers 210-07 and 210-15, respectively.

The terms“treatment”,“treating” and the like are used herein to generally mean obtaining a desired pharmacologic and/or physiologic effect. The effect may be prophylactic in terms of completely or partially preventing a disease or symptom thereof and/or may be therapeutic in terms of a partial or complete cure for a disease and/or adverse effect attributable to the disease. The term“genetic modification”, as used herein, refers to the direct manipulation of an organism’s genes. It comprises a subset of techniques, including gene delivery, mutagenesis and gene-editing.

“Gene editing” refers to the process of changing the genetic information present in the genome of a cell. This gene-editing may be performed by

manipulating genomic DNA, resulting in a modification of the genetic information. Such gene-editing may or may not influence expression of the DNA that has been edited. Gene editing may e.g. be achieved by using CRISPR-Cas technology or zinc finger nuclease technology.

The term“globular proteins”, as used herein, refers to proteins with a spherical-like geometry and includes serum globulins and albumins. Albumins and their fragments are very suitable as antigen- specific peptide. Albumins are water- soluble proteins and are found in all animals, including humans. The albumin or its fragment can be obtained from any source, preferably chicken. More preferably, a subset of albumins such as ovalbumins, or fragments thereof, are used. The ovalbumin or its fragments can be obtained from any source, preferably from chickens. More preferably, an ovalbumin fragment is a peptide comprising an epitope recognizable by major histocompatibility (MHC) molecules, e.g. a class I MHC molecule, H-2Kb. Most preferably, the OVA257-264 peptide fragment, is used. OVA257-264 peptide is a class I (Kb)-restricted peptide epitope of ovalbumin, and is presented by the class I MHC molecule, H-2Kb. The OVA257-264 peptide has the following amino acid sequence: N-terminal-SIINFEKL, with a molecular mass of 963.2 kDa and may be commercially obtained from Eurogentec (Liege, Belgium).

The term“fragment”, as used herein, should be understood to comprise any subset of amino acids of a protein with a length of at least three amino acids, preferably at least five amino acids, which comprise an antigen function. It is also intended that a fragment according to the invention can include conservative or non-conservative amino acid substitutions that do not substantially alter its biologic activity.

The term“molecular targets”, as used herein, refers to any compound or (molecular) structure originating from a living organism with which some other entity can interact on a molecular level, resulting in any change in the molecular target, e.g. its behavior or its function. Common molecular targets include nucleic acids and proteins, specifically proteins that have a receptor or a ligand-binding function.

Reference throughout this specification to“one embodiment” or“an embodiment” means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment envisaged herein. Thus, appearances of the phrases“in one embodiment” or“in an embodiment” in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures or characteristics may be combined in any suitable manner, as would be apparent to a person skilled in the art from this disclosure, in one or more embodiments. Furthermore, while some embodiments described herein include some but not other features included in other embodiments, combinations of features of different embodiments are also envisaged herein, and form different embodiments, as would be understood by those in the art.

It was demonstrated by M.Y. Balkhi et al. (M.Y. Balkhi et al., iScience 2, 105- 122, 2018) that in vitro T cells can be generated with decreased IL-2 and IFN _g production, and increased PD1, Tim3 and Lag3 expression. However, no loss of polyfunctionality has been shown, neither a reduced capacity to proliferate in vivo, nor the molecular gene expression pattern known to be associated with T cell exhaustion The described cytokine production changes by M.Y. Balkhi et al. are compatible with effector differentiation characterized by loss of IL-2 production. The decreased IL-2 and IFN _g production was assessed by measuring cytokine in supernatants without restimulation. The standard of exhaustion is set by restimulating T cells for 6h and determining their ability to produce cytokines and most importantly the ability of individual T cells to produce multiple cytokines simultaneously. The upregulation of inhibitory receptors, alone, is not sufficient evidence for exhaustion as these receptors are upregulated by T cell activation in the absence of exhaustion. Thus M.Y. Balkhi et al. have failed to provide critical evidence of exhaustion. Additionally, the in vitro generated T cells are not described as viable, i.e. preventing them from being cleared from the medium, thus reducing their applicability in in vitro experiments. It can thus not be concluded that partial or complete T cell exhaustion has been reached in the study of M.Y. Balkhi et al.. The current inventors now found that chronic antigen stimulation alone in the presence of the cytokines IL-7 and/or IL-15 results in the generation of in vitro viable, exhausted T cells.

The invention thus relates to a method for generating in vitro viable, exhausted T cells. Although the present experimental results focus on the generation of exhausted T cells originating from CD8 ⁺ T cells, the current invention is in no way limited to these particular cells. Other T cells that could be subject to T cell exhaustion include CD4 ⁺ T cells and naive T cells, or any other T cell known to the person skilled in the art. In one preferred embodiment the T cells are CD8 ⁺ T cells. The cells may be cultured in precence of antigen presenting cells. CD8 ⁺ T cells do not need antigen presenting cells to present peptides as they recognize peptides presented by major histocompatibility complex (MHC) I molecules which are found on the cell surface of all nucleated cells, including CD8 ⁺ T cells itself. However, other types of T cells may not be able to present peptides to themselves. For example, CD4 ⁺ T cells require other antigen presenting cells as they recognize peptides presented by MHC II molecules but lack MHC II expression. Culturing of such cells with antigen presenting cells is required for antigen presentation. Hence, in another preferred embodiment, the T cells are CD4 ⁺ T cells. If theT cells are CD4 ⁺ T cells, the repeated antigen stimulation is preferably in the presence of antigen presenting cells. Examples of antigen presenting cells are macrophages, dendritic cells, langerhans cells, and B- lymphocytes but any type of cells known in the art to be capable of presenting antigens can be used, as well as combinations of different antigen presenting cells. The antigen presenting cells and antigen can be added separately to the T cell culture. Alternatively, the antigen presenting cells are first incubated with the antigen and the antigen presenting cells are

subsequently added to the T cell culture.

One of the embodiments is the in vitro generation of exhausted cells, comprising culturing of T cells in a medium comprising at least IL-7 and/or IL-15, and repeated antigen stimulation of said T cells to exhaustion. In the present method, the concentration of T cells in culture can be varied. In practice, an optimal cell density or concentration of T cells is 10 ^˄ 5 cells/mL to 2-3 *10 ^˄ 6 cells/mL, more preferably 0.5* 10 ^˄ 6 cells/mL to 10 ^˄ 6 cells/mL (see tools. thermofisher. com/content/sfs/manuals/t-cell-activation-in-vitro. pdf).

T cells can be harvested from the spleen, lymph nodes or blood of any mammal such as a human, more preferably from a human or a rodent, most preferably a mouse. The cells are not limited to primary T cells but also cell lines are suitable, for example human or rodent, e.g. mouse, T cell lines. In one embodiment, when mice are used as a source of T cells, T cells of any mouse for which a T cell antigen is available, can be harvested. In a specific embodiment of the invention, an OT-I mouse (TCR transgenic ovalbumin specific CD8 ⁺ T-cell mouse; Hogquist et al., 1994. Cell 76:17-27) can be used for harvesting. The OT-I mice comprise two inserted transgenic genes, Tcra-V2 and Tcrb-V5, making it possible for the OT-I T cells to specifically recognize OVA257-264, which is presented by the Major Histocompatibility Complex I (MHC I).

The cell medium applied for culturing the T cells should at least contain the minimum: a buffer system, protein, trace elements, vitamins, inorganic salts, and energy sources, as is known to a person skilled in the art. Suitable media are known to a person skilled in the art. Preferably, RPMI 1640 is used as medium. In the medium, IL-7 and/or IL-15 are supplemented to the medium. Preferably, both IL-7 and IL-15 are supplemented. If only one of IL-7 or IL-15 is provided, it is supplemented to the medium in a high concentration, preferably 20-100 ng/mL, most preferably 25-80 ng/mL. If both IL-7 and IL-15 are added to the medium, a lower concentration of cytokines can be applied, such as a IL-7 concentration of 1- 50 ng/mL and a IL-15 concentration of 1-50 ng/mL, more preferably with a IL-7 concentration of 3-25 ng/mL and a IL-15 concentration of 3-25 ng/mL, most preferably a IL-7 concentration of 5-13 ng/mL and a IL-15 concentration of 5-13 ng/mL. The total concentration of IL-7 and IL-15 together in the medium

preferably is 1-100 ng/mL, more preferably 5-50 ng/mL, most preferably 10-25 ng/mL.

Antigens used in antigen stimulation can be any antigen that is recognized by T cells, and the chosen antigen should correspond to the chosen T cells, as will be clear to a person skilled in the art. More particularly, the T cell should be able to recognize the antigen for repeated antigen stimulation. Examples of T cell activating antigens include ovalbumin, viral or tumor proteins or peptides. Such antigens can be identified, for example, as detailed in published literature

(Blanchard, N., and Shastri, N. Curr Opin Immunol 20, 82-88, 2008) or using predictive algorithms and software (Andreatta, M and Nielsen, M, Bioinformatics 32, 511-517, 2016; Antunes D.A. et ah, Sci Rep 8, 4327, 2018; Liu G. et ah,

Gigascience 6, 1-11, 2017; Nielsen M. et ah, Protein Sci 12, 1007-1017, 2003). In addition, antigen stimulation may be provided by antibodies such as anti-CD3 antibodies and anti-CD28 antibodies. However, the method of the current invention is in no way limited to any of these antigens, other antigens and corresponding T-cell receptors can be found in literature and can be

advantageously used in the present invention.

The antigen preferably is an antigen- specific peptide. Said antigen-specific peptide preferably is selected from the group of globular proteins and fragments thereof, preferably ovalbumin and fragments thereof, most preferably OVA257-264 peptide.

The concentration of the antigen for repeated antigen stimulation can be varied. The concentration of the antigen can depend on the chosen antigen, and the particular reactivity of the T-cell receptor against the antigen. The concentration of the antigen preferably is from aboutl ng/mL to about 10 mg/mL, more preferably between 5 ng/mL - 5 mg/mL, most preferably between 10 ng/mL - 1 mg/mL.

Repeated antigen stimulation is carried out for at least one day, preferably for multiple days, preferably 2 to 15 days, more preferably 3 to 10 days, most preferably 5 days. It is known to the person skilled in the art, that the cells preferably are checked daily, antigen is refreshed daily. Feeding with fresh media and splitting should be implemented as necessary. Repeated antigen stimulation should be understood in such a way that the chosen antigen specific for the T-cell receptor is added repeatedly, preferably once daily or once every other day, to the medium of the cells.

T cells should be understood to be exhausted when three conditions are met:

1. when they cease to produce multiple cytokines and lose polyfunctionality; 2. when they express multiple inhibitory receptors simultaneously; 3. when the T cells exhibit decreased ability to proliferate upon restimulation. Different markers are used to evaluate these different parameters. The concentration of IL-2, IFN _g and/or TNFa may be measured by any technique known in the art, for example ELISA, antibody arrays, bead-based assays, Western blots. Said technique preferably comprises a fluorescence assay. In one embodiment, an indirect sandwich enzyme-linked immunosorbent assay (ELISA) can be performed. In this method, antibodies interact with the antigens present in the supernatant, followed by enzyme- or fluorescently labeled antibody detection via a colorimetric or fluorescent assay. Signal is measured via optical density or fluorescence and compared to a standard to determine the concentration of IL-2, IFN _g and/or TNFa. Results can be compared between different levels of exhausted T cells.

In another embodiment, that allows the assessment of polyfunctionality of T cells, cells are stimulated for 6h with antibody or peptide, in the presence of a Golgi plug and intracellular cytokine accumulation can be detected with fluorescently labeled antibodies against IL-2, IFN _g and/or TNFa. Cell may be analyzed on a flow cytometer and the simultaneous production of intracellular cytokines can be measured by measuring fluorescence emission at different wavelengths.

Cell cycle arrest may be assessed by evaluating the proliferation capabilities in vivo. Other methods to detect cell cycle arrest comprise flow cytometric analysis, measurement of cellular DNA content at a single time point, identification of proliferation-associated proteins and/or nuclear proliferation antigens, long incubation with thymidine analogs followed by said analogs’ detection, and any other commercially available kit, such as the“cell cycle assay kit” from Abeam and the“Cell Cycle Kit” from BD Pharmingen™.

In a particular embodiment of cell cycle arrest detection, cells are harvested, washed and resuspended in a buffer. The cells can then be fixed, for example with chilled ethanol and incubated at -20°C. The cells are again washed and a fluorescently tagged anti-nuclear proliferation antibody, such as for example an anti-Ki67 antobody and/or an anti-proliferating cell nuclear antigen (PCNA) antibody, is added. After an incubation time, the fluorescent signal is detected via flow cytometric analysis. Cells with higher fluorescent signals represent

proliferating cells, which are not in cell cycle arrest.

In another embodiment of cell cycle arrest assessment, cells are transferred into an animal, preferably a rodent. The animal can be vaccinated with a protein or peptides that are recognized by the T cells. Alternatively, animals can be vaccinated or infected with a vector with a DNA or RNA sequence that encodes for the cognate antigen of the T cell, such as a viral vector, can be introduced after a incubation period, for example after 3 hours. In the effector phase, donor cells can be counted with any cell counter, automatically or manually, known to a person skilled in the art. Proliferating T cells will result in a higher number of donor cells, when compared to exhausted, non proliferating T cells.

Resistance to reactivation can be determined from an increased expression of inhibitory receptors, as these receptors inhibit antigen stimulation. Expression of inhibitory receptors, such as PD-1, Tim-3a, Lag-3, CD160, CD244,TIGIT and others, is observed to be increased, preferably by a fluorescence assay. In one embodiment, cells are harvested and inhibitory receptors on the surface of the cells are stained with fluorescently-tagged antibodies. After a washing step, the antibodies can be detected in a fluorescent flow cytometric assay. The signal scales with the amount of receptors present on the surfaces of the cells. A high level of inhibitory receptors indicates exhaustion. The co-expression of multiple inhibitory receptors by the same cell reported with this assay is further evidence of

exhaustion.

One of the embodiments of this invention is a method for testing treatment to prevent, alleviate or accelerate exhaustion of T cells, comprising culturing T cells in a medium comprising at least IL-7 and/or IL-15, and repeated antigen stimulation of said T cells to generate exhausted T cells, wherein said T cells are treated before, during and/or after the culturing and/or the stimulation. Currently, no adequate in vitro system is available for researching T cell exhaustion, which has an extremely important role in both cancer and chronic infections. By eliminating all non- necessary factors and only having an environment with the minimally necessary cytokines and medium, methods to alter, such as to prevent, alleviate or accelerate, the exhaustion behavior of T cells can be tested.

In an embodiment, T cells are treated by administration of a chemical compound, preferably a small molecule, a peptide, a protein (including antibodies), a nucleotide and/or a carbohydrate, more preferably a peptide or a protein

(including antibodies). In such a case T cells are harvested from an animal, preferably a rodent, and cultured with IL-7 and/or IL-15, an antigen which can induce repeated antigen stimulation, and a chemical compound that is to be tested. The T cells are cultured over multiple days, preferably while daily adding or refreshing the antigen and the tested chemical compound. The cell medium with IL-7 and/or IL-15 can be refreshed when necessary. Subsequently, the level of exhaustion of the T cells can be evaluated, together with cell survivability. Cell survivability can for example be evaluated with a live/death assay. The latter one may also be used to evaluate the toxicity of the tested chemical compound. If a notable difference in exhaustion levels is detected while cell survivability is not altered, a chemical compound could be considered of particular interest for the treatment of cancer and/or chronic infections, especially as an additional treatment next to an anticancer treatment, antiviral or antibacterial treatment.

In another embodiment, T cells are treated with genetic modification, preferably by gene editing, most preferably by CRISPR-Cas technology or zinc finger nuclease technology. CRISPR-Cas technology, or Clustered Regulatory Interspaced Short Palindromic Repeats/ CRISPR-associated (Cas) system is nowadays the most successful and most widespread way of gene editing. In the Type II CRISPR/Cas system, short segments of foreign DNA, termed“spacers” are integrated within the CRISPR genomic loci and transcribed and processed into short CRISPR RNA (crRNAs). These crRNAs anneal to trans-activating crRNAs (tracrRNAs) and direct sequence- specific cleavage and silencing of pathogenic DNA by Cas proteins. Cas9 (CRISPR associated protein 9) is an RNA-guided DNA endonuclease enzyme. Cas9 has gained traction in recent years because it can cleave nearly any sequence complementary to the crRNA. The target specificity of Cas9 stems from the crRNA:DNA complementarity and not on modifications to the protein itself (like TALENs and Zinc-fingers). Hence, engineering Cas9 to target non- self DNA is straightforward. While native Cas9 requires a so-called guide RNA composed of the two disparate RNAs that associate to make the guide - the

CRISPR RNA (crRNA), and the trans-activating RNA (tracrRNA) - Cas9 targeting has been simplified through the engineering of a chimeric single guide RNA.

Scientists have suggested that Cas9-based gene alterations may be capable of editing the genomes of entire populations of organisms. In 2015, scientists used Cas9 to modify the genome of human embryos for the first time and in recent years the system has become immensely popular and widespread for gene editing of all kinds of eukaryotic cells. Next to Cas9 also other enzymes, such as Cpfl and other Cas enzymes would be applicable in the present invention. In an embodiment, CRISPR-Cas technology would be employed to screen for identification of genes involved in T cell exhaustion. For this, T cells preferably are provided in an arrayed format, for example a multi well plate such as a 96 well plate, 192 well plate of 384 well plate. Each well may be provided with a specific CRISPR-Cas combination that targets a specific gene. Screening each well for the ability of the T cells to generate exhausted T cells will result in the identification of one or more genes that play a role in the process of exhaustion of T cells.

Another gene-editing method is Zinc finger nuclease. In this technique, DNA sequences are specifically targeted with artificial zinc fingers. The sequences are cleaved at a specific site. This can be performed by fusing a zinc finger DNA- binding domain, for recognition of the DNA sequence, to a DNA-cleavage domain, which will be responsible for the DNA cleavage. After cleavage, foreign DNA could be introduced, for example with a plasmid vector. In one embodiment of the present invention, zinc finger nucleases could be generated that are able to introduce a gene of a molecular target. In a next step, the generated zinc finger nucleases are added to T cells and the DNA is cleaved in order to allow a new gene to enter the DNA of the T cell. Afterwards, the T cells can be exhausted and the exhaustion levels can be evaluated to determine the impact of the introduced gene.

In another embodiment, gene expression can be altered by knock in experiments employing, for example, DNA or RNA transfection or viral infection for example adenovirus, retroviruses or lentiviruses. These approaches can be used to deliver DNA or RNA or proteins that decrease or increase the expression of a gene or inhibit its activity. T cells can be transfected or infected and then subjected to the exhaustion method to identify genes that can prevent, stimulate, or revert exhaustion. T cells subjected to in vitro exhaustion can be tested in vitro or in vivo for the restoration of important functional characteristics and the decrease in inhibitory receptor expression.

In an embodiment, the T cells are evaluated in vivo, comprising the following steps: in vitro generation of exhausted T cells, transfer of the aforementioned in vitro-generated exhausted T cells into an animal, stimulation of the animal with the antigen, and detection of exhaustion markers. The transfer of the in vitro- generated exhausted T cells into an animal, is preferably carried out intravenously. By evaluating the T cells in vivo, a further assessment can be made regarding the level of exhaustion of the T cells. In one embodiment, live exhausted T cells can be transferred into an animal, preferably a rodent, most preferably a mouse. After sufficient time the animal is anesthetized and infected with a vector such as a viral vector carryinga targeting sequence of an antigen that is exclusively recognized by the exogeneous T cells, but not by the endogenous host mouse T cells. After an incubation period, the total amount of exogenous T cells may be determined, for example by counting with a cell counter. The number of exogenous T cells can be determined, for example by fluorescently labeled antibodies and flow cytometry, for example as described in Hope et al., 2017. Front Immunol 8:1696. doi:

10.3389/fimmu.2017.01696 and Gracias et ah, 2013. Nat Immunol 6: 593-602, and compared between samples of tested, exhausted and non-exhausted T cells. A reduction in T cells compared to non-exhausted T cells indicates a reduction in cell proliferation, and thus a higher level of T cell exhaustion.

In addition, a treatment such as a small compound, protein, or genetic modification that is identified as preventing, restoring or accelerating exhaustion of T cells in the in vitro methods of the invention, preferably is tested in vivo. For this, animals are infected, for example acute or chronic infection with a bacterium or virus, or animals carrying a tumor that can be targeted by the exhausted T cells, can be tested for their ability to enhance efficacy of the exhausted cells to control or clear said bacterial or viral infections or said tumor.

An embodiment of the current invention is a method for identifying molecular targets to prevent, restore or accelerate T cell exhaustion, comprising culturing T cells in a medium comprising at least IL-7 and/or IL-15; repeated antigen stimulation of said T cells to generate exhausted T cells; treating said T cells with a treatment that is able to prevent, restore or accelerate T cell exhaustion; and comparing the treated T cells with untreated exhausted and non-exhausted T cells, to find relevant changes in said T cells. An important advantage of this approach with using in vitro-generated exhausted cells, is the removal of non-relevant influences originating from environmental signals that are typically introduced in an in vivo-model. By using in vitro-generated exhausted cells with only the necessary cytokines, it is thus possible to identify molecular targets that are exclusively due to exhaustion.

By researching treatments to prevent, restore or accelerate exhaustion of T cells, molecular markers could be identified that are required for T cell exhaustion. New treatments for cancer and chronic infections can be developed based on these molecular markers. In an embodiment of this method to identify molecular targets, an inhibitor of a specific molecule and/or receptor is added to the T cells in culture, either before, during and/or after inducing T cell exhaustion. Subsequently, the level of exhaustion of the T cells can be evaluated, together with cell survivability. Cell survivability can for example be evaluated with a live/death assay. The latter one is necessary to evaluate the toxicity of the tested chemical compound. If a notable difference in exhaustion levels is detected while cell survivability stays stabile, the specific molecule and/or receptor could be considered of particular interest.

Another embodiment of the current invention is a cell culture comprising viable exhausted T cells. Further provided is a cell culture comprising viable exhausted T cells obtained by any method according to the invention. As shown in the Examples, no T cells obtained with a method of the invention make IL-2 and only a small percentage makes TNFa. Nearly all cells (> 95%) express Tox and PD- 1 ((> 95%) and the vast majority (> 80%) expresses multiple inhibitory receptors. Hence, in a preferred embodiment, in the cell culture:

- more than 80%, preferably more than 90%, more preferably more than 95%, of the cells express Tox,

- more than 75%, preferably more than 79%, of the cells express PD-1,

- more than 70% of the cells express two or more inhibitory receptors, preferably two or more inhibitory receptors selected from the group consisting of PD-1, Tim-3, TIGIT, LAG 3, CD 160 and CD244,

- at most 10%, preferably at most 5%, more preferably at most 2%, of the cells produce IL-2, and/or

- at most 10%, more preferably at most 5%, more preferably at most 2%, of the cells produce IFN- _U and TNF-a.

Features may be described herein as part of the same or separate aspects or embodiments of the present invention for the purpose of clarity and a concise description. It will be appreciated by the skilled person that the scope of the invention may include embodiments having combinations of all or some of the features described herein as part of the same or separate embodiments.

The invention will be explained in more detail in the following, non-limiting examples.

Examples

Example 1

Material and Methods

Mice. OT-I CD45.1+ mice on the C57BL6/J background were generated by backcrossing C57BL/6 Tg (TcraTcrb)1100Mjb/J (OT-I) with B6.SJL-Ptprca

Pepcb/BoyJ (CD45.1+) mice (both from the Jackson Laboratory). C57BL/6J mice and OT-I mice were housed in a certified barrier facility at Erasmus University Medical Center. Animal work was performed under Project Proposal

(AVD101002015179) by the animal welfare body (AWB) of the Instantie voor Dierwenwelzijn (IvD). All animal experiments were conducted in compliance with the Netherlands’ government laws of the Centrale Commissie Dierproeven (CCD).

8-12 weeks old Female mice were anesthetized with 2.5% isoflurane gas and were infected intranasally with Influenza virus strain A/WSN/33-expressing OVA(257-264) (WSN-OVA, a gift from D. Topham, University of Rochester Medical Center).

Repeated antigen stimulation in vitro. Firstly, spleen was harvested from OT-1 mouse and the tissue was processed to get single cell suspension. Then a CD8 purification kit (EasySep; Stemcell Technologies) was used to isolate CD8+ T cells from these splenocytes. 0.5 *10 ^˄ 6 cells/ml of the Purified CD8+ T cells were cultured in complete media presented with cytokines IL-15 (5ng/ml, Peprotech, Cat 210-15) and IL-7 (5ng/ml, Peprotech, Cat 210-07) with or without lOng/ml OVA257- 264 peptide (Anaspec, Cat AS-60193).

For single peptide stimulation, cells were cultured in the presence of a peptide for 48 hours, before the peptide was removed by washing the cells two times with 10% RPMI medium. For the remaining 3 days, the single peptide stimulated cells were cultured in the media with cytokines. For repeated peptide stimulation, lOng/ml OVA257-264 peptide was added daily for five days. Repeated peptide stimulated cells were washed on day 2 to allow for comparable culture conditions. No peptide stimulation control cells were cultivated in media with cytokines but without adding a peptide. Cells from all these three conditions were checked daily, feeding with fresh media or splitting were implemented as necessary.

On day five, cells were harvested and counted by using an automated counting system (Countess, life technologies). DAPI Viability dye (Beckman Coulter, Cat B30437) and Acridine Orange (Biotium, Cat 40039) were used to dye the cell to distinguish live and dead cells.

In vitro killing assay. AE17 cells were maintained in RPMI 1640 supplemented with 10% FBS (Gibco), 100 units/mL Penicillin/Streptomycin (Life Technologies), 2 mM L-glutamine (Life Technologies), 0.05 mM 2-mercaptoethanol (Sigma), and were cultured at 37°C in 5% CO2. AE17 cells were pulsed with 1 mg/ml OVA(257-264) Peptide (Anaspec Cat AS-60193) for 1 hour and then the cells were washed thoroughly before they were labelled with the CellTrace™ Far Red fluorescent dye (ThermoFisher Scientific Cat C34564/15396613). Un-pulsed cells were not labeled. A 1:1 mix of peptide pulsed and un-pulsed AE17 cells (105 each) were mixed and different amounts of T cells (Effector: Target ratios: 3:1, 1:1, 0.3:1) were added. The cells were harvested after 16 hours, the ratio of labeled and unlabeled tumor cell were detected by flow cytometry.

Flow cytometry. To investigate whether single or multiple peptide stimulated CD8+ T cells show different functions, these cells were first

phenotypically characterized by flow cytometry. Cells were stained immediately after harvesting. Surface expression of inhibitory receptors and ligand antibodies were: CD8-ef450 (53-6.7, eBioscience); Lag3-APC (C9B7W, BD); PD-l-APC-Cy7 (19F.1A12, Biolegend); CD244-PE (2B4, BD; eBio244F4, eBioscience); Tim3-PE- Cy7 (RMT3-23, Invitrogen); CD160-CF594 (CNX46-3, BD); TIGIT-FITC (GIGD7, eBioscience); PD-L1-BV711 (MIH5, BD). Other CD8 function differentiation and development related markers were: CD44-BV785 (IM7, BD) ; CD25-APC-Cy7 (PC61, BD); CD127-PerCP-Cy5.5 (A7R34, Biolegend); KLRG1-PE-Dazzle594 (2F1, Biolegend); CD28-FITC (E18, Biolegend). Intracellular expression of transcription factors was determined with Tbet-PE-Cy7 (4B10, Biolegend); TCF1-APC (C63D9, Cell Signaling); EOMES-PE-eF610 (Danllmag, eBioscience); Tox-PE (TXRX10, eBioscience). To exclude dead cells, Annexin V (APC, BD; Cy5.5, BD; PerCP-Cy5.5, BD) stain was included in all the stains and 2.5 mM CaCl ₂ was added to all solutions.

On day 5 of culture, cells were harvested and immediately stained for surface and intracellular antigens. For surface staining, cells were washed with FACS wash (3% FBS in Hanks' balanced salt solution (HBSS)) + 2.5 mM CaCl ₂ and incubated with 20 mL mix of the pre-determined optimal concentrations of the fluorochrome-conjugated monoclonal antibodies on ice in the dark for 20 minutes. The cells were again washed one time with FACS wash and fixed with 1% PFA + 2.5 mM CaC12. For the transcription factor staining, the cells were first stained for surface antigens as described above. Following the washing step, cells were fixed with FoxP3 Fixation Buffer (005523, eBioscience) for 60mins in the dark on ice and then washed with Perm/Wash buffer (008333, eBioscience) + 2.5 mM CaCl ₂ and stained with a mix of antibodies against transcription factors for 1 hour in the dark at 4°C. The cells were washed then twice with Perm/Wash buffer + 2.5 mM CaCl ₂ and fixed with 1% PFA + 2.5 mM CaCl ₂ . Intracellular expression of transcription factors anti-Tbet-PE-Cy7 (4B10, Biolegend); anti-TCFl-APC (C63D9, Cell

Signaling); anti-EOMES-PE-eF610 (Danllmag, eBioscience); anti-Tox-PE

(TXRX10, eBioscience) were measured; Appropriate isotype controls were included for staining of transcription factors.

To analyze cytokine production, cells were re-stimulated with the OVA N- terminus SIINFEKL peptide for 6 hours at 37°C, 5% CO2 in the presence of GolgiPlug (BD Biosciences) and anti-CD107a-APC-Cy7 antibodies (clone ID4B, Biolegend). Cells were then stained with surface marker antibodies (anti-CD8, anti-PD-1, anti-CD44) as described above. After washing with FACS wash, cells were fixed with IC Fixation Buffer (88-8824, eBioscience) overnight at 4°C and then washed with Perm/Wash buffer and stained for intracellular cytokines for 45 min in the dark at 4°C. After staining, the cells were washed twice with

Perm/Wash buffer and fixed with 1% PFA. Cytokines were measured with anti- IFN- _U -APC (XMG1.2, eBioscience); anti-TNF-a-AF488 (MP6-XT22, eBioscience); anti-IL-2-PE (JES6-5H4, eBioscience); anti-GranzymB-PE-Cy7 (NGZB,

eBioscience); or the appropriate isotype controls were used for intracellular stains. After staining, cells were washed twice with Perm/Wash buffer (eBioscience) and fixed with 1% PFA.

Cells were run on a LSRFortessa (BD Biosciences) and at least 200,000 events were collected. Data was then analyzed with FlowJo software (Version 9.9.4, Treestar, Ashland, OR, USA).

In vivo Influenza model. To analyze whether the in vitro stimulated cells have proliferative defects in vivo, in vitro generated cells were adoptively transferred into wild type mice. We adoptively transferred CD45.1+ OT-I cells from in vitro cultured cells that were first were sorted on a FACSAria III (BD

Biosciences) using fluorochrome-conjugated Annexin V and anti-CD8 antibodies. After sorting, 10,000 cells were intravenously transferred into 8—12 weeks old CD45.2+ C57BL/6J wild-type recipient mice. Naive CD8+ T cells were also freshly isolated from spleens of an OT-1 mouse and adoptively transferred. The mice were then anesthetized with 2.5% isoflurane gas and were infected intranasally with influenza virus strain A/WSN/33-expressing OVA(257— 264) (WSN-OVA, a gift from Dr. D. Topham, University of Rochester Medical Center). Three hours later, donor OT-I cells were adoptively transferred. Body weight was measured daily to track the influenza infection.

Ten days post infection, the lung, spleen and mediastinal lymph nodes were harvested and single cell suspensions were obtained after processing the tissues.

To prepare single cell suspensions, lungs were digested for 2 h at 37°C with 3.0 mg/ml collagenase A and 0.15 mg/ml DNase I (Roche) in RPMI 1640 containing 5% heat-inactivated FBS, 2 mM 1-glutamine, 100 IU/ml penicillin, 100 mg/ml streptomycin. Digested lung was then run through a 40-pm cell strainer (Falcon ) and washed in the same media as above. PE-conjugated tetramers of H-2b major histocompatibility complex class I loaded with OVA(257-264) was used to identified the antigen-specific CTLs in the lung. Part of the lung tissue was frozen in TRIzol LS reagent (Life Technologies), RNA was isolated (described below) and the viral load was determined by qPCR.

RNA sequencing. To compare the transcription expression level among those cells, RNA sequencing was performed. On day 5, 1*10 ^˄ 6 live CD8 ⁺ cells were sorted from the three different cultured cells and immediately lysed with TRIzol LS reagent (Life Technologies) and stored at -80°C. RNA was extracted according to manufacture’s instructions and a bioanalyzer (Agilent) was used to determine the integrity of the extracted RNA. Barcoded sequencing libraries were generated using a KAPA RNA Hyper+RiboErase HMR (Roche Diagnostics). Library quality was assessed with the bioanalyzer and Kapa qPCR was performed before cluster generation and 100-bp paired-end sequencing. Libraries were sequenced with HTSeq2000 Illumina technology.

Analysis of differential transcript abundance. Four independent biological replicates were analyzed for each condition. The quality of the sequencing was checked by using Fast QC software (see

bioinformatics.babraham.ac.uk/projects/fastqc/). The demultiplexed fastq files were aligned using STAR software (v.2.5.3e) using default strings and Mus_musculus GRCm38 as alignment reference. Then Binary Alignment Map (BAM) files were annotated using Feature Count software (vl.6.1) to obtain the annotated files (count files), annotation reference was genecode.vM15.gtf.

Count data was preprocessed to remove the very low expression genes, then, an rlog transformation (DESeq2 R package (vl.22.2) was performed with the clean count files. This transformation was exclusively for visualization and comparison purpose which included correlation and clustering analysis of the data (heatmaps and PCA plots). For differential expression analysis DESeq2 package was used in the raw data (not rlog transformed).

The list of differential expressed genes was used to perform IPA (Ingenuity Pathway analysis) (IPA, Qiagen, USA version) to further discern which pathways are involved in CD8 ⁺ T cell exhaustion process. Pathway enrichment P values (Fisher’s exact test) and activation Z-scores were calculated by IPA and used to rank the significant pathways.

For gene set enrichment analysis (GSEA), our gene-list was the list of the significantly differentiated genes obtained from the comparison between single peptide stimulated and multiple peptide stimulated. The GSEA Desktop

Application (v2.2.1) was used to run this analysis. A separated list of up regulated and downregulated genes in cell exhaustion were used as gene-sets for GSEA analysis. Two CTL exhaustion gene-sets (up and downregulated) were downloaded from the Broad Institute (see Molecular Signatures Database from the Broad Institute). Normalized Enrichment Scores were calculated using the function GseaPreranked.

DNA methylation profiling detection. DNA methylation profiling was done as previously described by the MeD-seq method. For MeD-seq sample preparation LpnPI (New England Biolabs) digestions were carried out on DNA samples according to manufacturer’s protocol. Reactions contained 50 ng and digestion took place overnight in the absence of enzyme activators. Digests of genomic DNA with LpnPI resulted in snippets of 32 bp around the fully- methylated recognition site that contains CpG. The DNA concentration was determined by the Quant-iT™ High- Sensitivity assay (Life Technologies) and 50 ng dsDNA was prepared using the ThruPlex DNA-seq 96D kit (Takara). Twenty microliters of amplified end product were purified on a Pippin HT system with 3% agarose gel cassettes (Sage Science). Stem-loop adapters were blunt end ligated to repaired input DNA and amplified (4 +10 cycles) to include dual indexed barcodes using a high fidelity polymerase to yield an indexed Illumina NGS library. Multiplexed samples were sequenced on Illumina HiSeq2500 systems for single read of 50 base pairs according to manufacturer’s instructions. Dual indexed samples were demultiplexed using bcl2fastq software (Illumina).

MeD-seq data processing was carried out using specifically created scripts in Python version 2.7.5. Raw fastq files were subjected to Illumina adaptor trimming and reads were filtered based on LpnPI restriction site occurrence between 13-17 bp from either 5’ or 3’ end of the read. Reads that passed the filter were mapped to mmlO using bowtie2.1.0. Multiple and unique mapped reads were used to assign read count scores to each individual LpnPI site in the mmlO genome. BAM files were generated using SAMtools for visualization. Gene and CpG island

annotations were downloaded from UCSC (MM10). Genome wide individual LpnPI site scores were used to generate read count scores for the following annotated regions: transcription start site (TSS) (1 kb before and 1 kb after), CpG islands and genebody (1 kb after TSS till TES).

MeD-seq data analysis was carried out in Python 2.7.5. DMR detection was performed between two datasets containing the regions of interest (TSS, genebody or CpG islands) using the Chi-Squared test on read counts. Significance was called by either Bonferroni or FDR using the Benjamini-Hochberg procedure. Differently methylated regions were used for unsupervised hierarchical clustering, the Z-score of the read counts was used for normalization and is also shown in the heatmaps.

In addition, a genome wide sliding window was used to detect sequentially differentially methylated LpnPI sites. Statistical significance was called between LpnPI sites in predetermined groups using the Chi-squared test. Neighbouring significantly called LpnPI sites were binned and reported, DMR threshold was set at a minimum of ten LpnPI sites, a minimum size of 100 bp and either a twofold or fivefold change in read counts. Overlap of genome wide detected DMRs was reported for TSS, CpG island and gene body regions

Statistical analysis. Statistical analyses were performed using Prism software (GraphPad Prism5 for Windows, Version 5.04). The normality of data distribution was assessed using the Shapiro-Wilk normality test. Homogeneity of variance was tested with Bartlett’s test. When data were normally distributed and group variances were equal, an ANOVA with Tukey's Multiple Comparison Test was performed. When data were normally distributed but group variances were unequal, a Student’s t test with Welch’s correction was performed. If data were not normally distributed a Wilcoxon signed rank test or a Mann— Whitney U test was performed. P values equal or lower than 0.05 were considered statistically significant with the numbers of stars in the figures indicating the p value: * = P ^ 0.05, ** = P £ 0.01, and *** = P £ 0.001.

Results

Repeat stimulation with cognate peptide to induce CTL exhaustion

Because of the critical role sustained antigen stimulation plays in driving CD8+ T cells to exhaustion, we utilized repeat peptide stimulations of OVA(257-264) - specific TCR transgenic OT-I cells to induce CTL exhaustion in vitro. To induce exhaustion, CD8+ T cells purified from OT-I mice were stimulated daily for five days with 10ng/ml OVA(257-264) peptide in the presence of IL-15 and IL-7 (Repeat peptide stimulation, Figure 1A). As controls, cells were either left unstimulated or stimulated only once with OVA(257-264) peptide for 48 hours and then washed and cultured without peptide for an additional 3 days. All cells were cultured with IL-7 and IL-15. On day 5, cells were either harvested for analysis or in some instances cells were cultured an additional 3 days in the presence of IL-7 and IL-15 without any additional peptide stimulation (Figure 1A).

An important feature of this in vitro exhaustion system is that it rapidly yields large numbers of exhausted cells within 5 days (Figure IB). Single peptide stimulations yielded comparable numbers while unstimulated culture numbers remained largely unchanged. The expansion in the numbers of exhausted T cells represents a ~9-fold increase from the T cell numbers seeded. Thus this in vitro exhaustion system can yield large numbers of exhausted CTL that can be used to further study and manipulate the pathways that drive exhaustion. Further, large numbers of exhausted CTL can also be obtained using this in vitro exhaustion system by culturing the cells in precence of IL-7 alone or IL-15 alone (Figure 1C). The total amount of viable exhausted cells yielded in the repeated peptide stimulated cultures was comparable for cells cultured with IL-7 or IL-15 on their own and cells cultured with a combination of both cytokines. However, the combination of both IL-7 and IL-15 yielded more viable exhausted cells in the single peptide stimulation cultures

Repeated peptide stimulation reduces cytokine production. Since hierarchical loss of the capability to produce cytokines is one of the most critical characteristics of exhausted CD8+ T cells, we first detected the cytokine release upon re-stimulation by flow cytometry. After five days of culture, the cells were harvested and stimulated by OVA-peptide with a Golgi plug for 6 hours. Strikingly, ~70% of repeatedly stimulated cells produced no cytokines. More than 40 percent of the cells without peptide stimulation produced Interferon-g (IFN-g) upon re- stimulation. Single peptide stimulated cells also produced IFN-g (57.95±3.82%). In contrast, repeated peptide stimulated cells had impaired function to release IFN-g, only 26.92 percent of the cells were IFN-g positive after re-stimulation (Figure 2A). In addition, there were much less TNF-a-producing cells in the repeated peptide stimulated cells (4.73±0.43%) than in single peptide stimulated cells (44.48±4.25%) and without peptide stimulated cells (71.14±4.08%) (Figure 2A). Interestingly, we could barely detect IL-2-producing cells in the repeated peptide stimulated cells after peptide re-stimulation. While the cells from the other two conditions did release IL-2, although IL-2 positive proportion was lower in single peptide stimulated cells (36.92±3.84) than in no peptide stimulated cells (60.34±3.84) (Figure 2A). These functional results therefore demonstrated that repeated peptide stimulated cells impaired the cytokine producing functions in an order similar to the exhausted CD8 ⁺ T cells, IL-2 first and then TNF-a, IFN-g the last.

Deficiency of polyfunctionality of repeated peptide stimulated cells.

Because exhausted CD8 ⁺ T cells cannot produce multiple cytokines at the same time as memory cells do, we next measured the polyfunctionality of these different treated cells under peptide re-stimulation. To achieve this, percentage of single cytokine producing cells, double cytokine producing cells and triple positive cells were output from flowjo software (FlowJo, LLC) and visualized by using Simplified Presentation of Incredibly Complex Evaluations (SPICE). As presented, in the no peptide stimulated cells, one third of the cells could produce three cytokines at the same time. Additionally, around 10 percent of the cells were IFN-g and TNF-a positive, 18 percent were TNF-a and IL-2 positive, around 5 percent of the cells were IFN-g and IL-2 positive. 15 percent of these cells did not produce any of these three cytokines (Figure 2B). Similarly, 18 percent of the single peptide stimulated cells were triple positive for all the three cytokines, around 30 percent of the cells were double cytokine producers. About one quarter of these cells released none of these cytokines (Figure 2B). In contrast to these, 70 percent of the repeated peptide stimulated cells could not produce any cytokines after re-stimulation. Around one quarter of these cells produce only IFN-g, 4 percent and 1 percent of the cells wereIFN-g+ TNF-a+ and only TNF-a+, respectively (Figure 2B). Clearly, after five times of daily peptide stimulation, CD8 ⁺ T cells are losing their polyfunctionality. They could not react on peptide re- stimulation and therefore they were in a functional deficient state.

Degranulation is an important step for the CD8+ T cell cytotoxicity. We therefore analyzed the degranulation marker CD 107a and Granzyme B (GzmB) expression in our cells. When cells were analyzed without peptide re-stimulation, cells either unstimulated or single peptide stimulated were negative for CD107a and GzmB (Figure 3A and 3B), while repeat peptide stimulated cells showed a significantly higher level of both markers (Figure 3A and 3B). When cells were re-stimulated with OVA(257-264) peptide, cells cultivated in the presence of no peptide and single peptide stimulated cells upregulated CD107a (Figure 3A). In contrast, repeat peptide stimulated cells failed to increase CD107a expression after re-stimulation, indicating that these cells are not able to degranulate further upon re- stimulation. Granzyme B was increased in repeat peptide stimulated cells in agreement with the terminal differentiation of exhausted cells (Fig 3B). No increase of GzmB was detected after 6 hours of re- stimulation for any of the conditions analyzed (Figure 3B). Repeat peptide stimulated cells exhibited reduced cytotoxic capacity against OVA(257-264) peptide-loaded tumor cells (Fig 3C). These findings clearly show that after repeat stimulation in vitro, CTL lose their ability to make cytokines, have reduced polyfunctionality, cannot degranulate further, have increased GzmB and reduced cytotoxicity. All of these features are characteristics of the dysfunctional state of in vivo exhausted CTL.

Multiple inhibitory receptors upregulated after repeated peptide stimulation. Having demonstrated that repeated peptide stimulated CD8 ⁺ T cells were deficient in cytokine production upon re-stimulation, we sought to determine the inhibitory receptors expressing on these cells. Normally, an increase in inhibitory receptor expression means that the cells are more exhausted and there are more inhibitory pathways that can be researched with this in vitro model. After harvesting the cells on day five, surface staining was used to measure the inhibitory receptors expressed on the differently treated cells. As expected, inhibitory receptors were barely expressed on the cells without peptide stimulation and single peptide stimulated cells (Figure 4A). Unsurprisingly, roughly 20 percent of the cells were PD-1 positive cells in the single peptide stimulated cells. In contrast, in vitro exhaustion (multiple peptide) cells are roughly 80% positive for PD-1 (Figure 4A). All other inhibitory receptors are increased on in vitro exhausted (multiple peptide) cells (Figure 4A). Very few no peptide or single peptide stimulated cells expressed more than one inhibitory receptor (Figure 4B). In contrast the majority of cells submitted to in vitro exhaustion (multiple peptide) expressed more than 2 inhibitory receptors (Figure 4B). We next examined the simultaneous co-expression of multiple inhibitory receptors (PD-1, Lag3, CD160 and CD244) using SPICE (Fig 4C). Repeat peptide stimulated cells were 39% double positive for the inhibitory receptors and another 39% of the cells co-expressed three of these inhibitory receptors simultaneously. Furthermore, 11% of the cells expressed all 4 inhibitory receptors (Fig 4C). In contrast, very few of the no peptide and single peptide stimulated cells expressed two or more inhibitory receptors (Fig 4C).

To exclude that differences in inhibitory receptor expression were due to a different activation status, all the cells were re-stimulated for 6 hours after harvesting them on day 5. Although reactivation induces a slight upregulation of some inhibitory receptors on the no peptide and single peptide cultures, they still remain much lower than the repeat peptide stimulated cultures. The exception was CD 160 which was upregulated to similar levels in all cells. Overall, these findings confirm that multiple inhibitory receptors are expressed on the repeat peptide stimulated cells.

Expression of transcription factors is altered in repeated peptide receptor expression indicated stimulated cells. The in vitro analysis of cytokine production and inhibitory receptors showed that the repeated stimulated cells represent exhausted CTLs. Previous studies have reported that exhausted CTL express and utilize different transcription factors (TF), when compared to effector and memory cells (Miller et al., Nat Immunol 20, 326-336, 2019; Xu L. et ah, Nat Immunol 16, 991-999, 2015). To characterize the expression pattern of TF in the in vitro exhausted cells, expression levels of four of the core TFs, T cell factor- 1 (Tcfl), Thymocyte selection-associated HMG box protein (Tox), T-box transcription factor 21 (T-bet), Eomesodermin (EOMES), were measured. Tcfl has been reported to play a critical role in identifying the subsets of exhausted T cells (Miller B.C. et ah, Nat Immunol 20, 326-336, 2019; Xu L. et ah, Nat Immunol 16, 991-999, 2015). Early exhausted or the progenitor exhausted cells maintained Tcfl expression, while terminally exhausted T cells downregulate its expression. In comparison to single peptide stimulation, repeated peptide stimulation induced down-regulation of Tcfl expression (Figure 5). Tox was reported to be increased expression on exhausted cells (Bengsch B. et ah, Immunity 48, 1029-1045, 2018; Miller B.C. et ah, Nat Immunol 20, 326-336, 2019), and was significantly upregulated in the repeat peptide stimulated cells in comparison to unstimulated and single peptide stimulated cells. T-bet and EOMES expression in exhausted CD8 T cells varies in different stages of differentiation and exhaustion (McLane L.M. et al., Annu Rev Immunol 37, 457-495, 2019; Paley M.A. et ah, Science 338, 1220-1225, 2012). After in vitro repeat peptide stimulation, T-bet was highly upregulated, while there was no detectable difference of EOMES expression from single peptide stimulated cells (Figure 5). The increased Tox and T-bet

accompanied by decreased Tcf-1 expression in repeat peptide stimulated CTL is compatible with the TF profile of in vivo exhausted CTL.

Repeated peptide stimulated cells decreased in vivo expansion capacity. To further prove that the repeated peptide stimulated CTLs are exhausted, their in vivo proliferative capacity was analyzed in a viral infection model. To examine the in vivo expansion capability of the differently treated cells, 10,000 of the sorted live CD8 ⁺ T cells were adoptively transferred into wild type mice on day 0. Three hours later, the mice were infected with WSN-OVA influenza virus intranasally. In the effector phase, ten days post infection, the number of donor cells in the infected lung were measured. We found a smaller proportion of repeated peptide stimulated cells, compared to no peptide and single time peptide stimulated cells, in the lung of the mice (22.9%, 67.9% and 66.4%, for repeat peptide, single peptide and no peptide respectively; see Figure 6a). The absolute donor cell numbers were consistent with the increased percentage of donor cells in total CD8 ⁺ T cells (3.18*10 ^˄ 6, 11.78*10 ^˄ 6 and 12.63*10 ^˄ 6, for repeat peptide, single peptide and no peptide respectively (Figure 6b). Although, the body weight loss and viral load were not significantly different among the mice that received differently treated cells, the mice that received the multiple times peptide stimulated cells possessed larger amount of endogenous OVA-specific CD8 ⁺ T cells in their lungs than the mice that received no peptide treated cells or single time peptide stimulated cells (0.63*10 ^˄ 6 versus 0.13*10 ^˄ 6 and 0.22*10 ^˄ 6), indicating that repeatedly stimulated cells were less fit to compete with the endogenous CD8+ T cell response (Figure 6c and d). These results together indicate that after multiple times of peptide stimulation, the cells decrease their expansion capacity when they were activated by the specific peptide again. Repeated peptide activated cells have transcriptional features of exhausted CTLs. To determine whether the differently treated cells also have distinct transcriptional profiling, we defined transcriptional states of the sorted live CD8 ⁺ T cells from the different cultivation conditions and compared them between each other. As presented in the principle component analysis (PCA), the samples from the same treatment were closely clustered, while different types of stimulation made the cells group distinctly in the PCA (Figure 7A). We identified 1196 genes with increased expression in repeated peptide stimulated cells, relative to the genes in single time peptide stimulated cells. Among the upregulated genes, multiple inhibitory receptors encoding genes, including PD-1 (pdcdl), Lag3, Tim3 (Havcr2), CD 160, TIGIT and CTLA4 were presented on the top of the list (Figure 7B). Unsurprisingly, the genes for the markers of the terminally differentiated effector cells, like GzmB, Gzmc, and prfl were also expressed significantly higher in the repeated peptide stimulated cells. The transcription factors EOMES and TCF7 were significantly downregulated in the repeated peptide stimulated cells, while Tox was found to be three times more expressed in these cells than in single peptide stimulated cells (Figure 7B). The above changes of single gene expression detected by flow cytometry were in line with the findings of RNA level modification. There were other exhaustion-related genes that were identified to be differentially expressed on repeated stimulated cells compared to single peptide stimulated cells(Figure 7B). Furthermore, transcription factors, which are associated with CTL exhaustion such as IRF4, NR4a and Batf were also upregulated in the repeat peptide stimulated cells in comparison to single peptide and no peptide stimulated cells (Figure 7B).

In order to assess differentially expressed genes by gene sets instead of single gene expression, gene set enrichment analysis was performed. The exhaustion upregulated gene sets were significantly enriched in the repeated peptide stimulated upregulated genes comparing to single peptide stimulated cells and no peptide stimulated cells. On the opposite, the gene sets that were reported to be downregulated in the exhausted cells were more enriched in the genes that down regulated in repeated peptide stimulated cells versus single peptide and no peptide stimulated cells. Ingenuity pathway analysis was utilized to further evaluate the pathway that is upregulated or downregulated in the chronic peptide stimulated cell comparing to the other two conditions. Obviously, the CTL exhaustion signaling pathway was the top pathway that positively related with repeated peptide stimulation of cells (Figure 7C). Simultaneously, Glycolysis I pathway was also significant positively related to the repeated peptide stimulated cells (Figure 7C).

Repeat peptide stimulation results in hyper-methylation of the Tcf7 transcriptional regulatory region

Changes in methylation of transcription regulatory region have been described in exhausted CTLs. In order to identify whether in vitro repeat peptide stimulated cells also possess similar epigenetic characteristics of exhausted CTLs, whole genome methylated DNA sequencing (MeD-seq) was performed on the sorted live CD8+ T cells. Stimulated T cells undergo distinct genome wide DNA methylation .changes depending on the type of treatment (Figure 8A). In comparison to unstimulated and single peptide stimulated cells, the transcriptional regulatory region of Pdcdl had significantly less DNA methylation in the repeat peptide stimulated cells (Figure 8B, left). When comparing the promotor methylation status (2kb region surrounding the TSS) of Tcf7, there was more DNA methylation detected in repeat peptide stimulated than in the single peptide stimulated cells or unstimulated cells (Figure 8B, center). Meanwhile, in the promotor region of the GzmB, significantly less DNA methylation was found in the repeat peptide stimulated cells than in the other cells (Figure 8B, right), which was in line with the higher expression of the protein upregulated in the repeat peptide stimulated cells (Figure 3B). Besides Pdcdl, none of the other inhibitory receptor genes’ regulatory or promotor regions were found to possess significant differences in methylation status, although some of them showed expected trends in DNA methylation changes at their TSS. Interestingly, the DNA methylation states of the cytokine genes IL-2, IFN-g and TNF-a were not significantly different. This indicated that other gene expression control mechanisms, like histone

modifications or transcription factor abundance, might regulate the differential expression of these cytokines. Overall, these findings indicate that the repeat peptide stimulated cells have distinct DNA methylation patterns and reveal that the downregulation of TCF1 expression is accompanied by increased promotor methylation of Tcf7.

To further confirm that DNA promotor methylation contributes to the

downregulation of TCF1 expression in exhausted cells we treated cells with a DNMT inhibitor for the last 3 days of culture and examined whether TCF1 expression was modified in the repeat peptide stimulation cultures. Indeed, treatment of these cells with DNMT inhibitor resulted in an increase in TCF1 levels in exhausted CTL (Figure 8C). This further supports that DNA methylation plays a role in silencing TCF1 as exhausted cells progress from“progenitor exhausted” to the“terminally exhausted” subpopulation.

DNMT inhibitor also reduced the expression of PD-1 and Tim3 in repeat peptide stimulated cells (Fig 9). Because PD-1 and Tim3 expression are controlled by promoter region DNA methylation, our findings suggest that DNMT inhibitor prevented exhaustion rather than reverted the exhaustion of T cells. These experiments suggest that the in vitro exhaustion system we describe can be used to test reagents that modulate T cell exhaustion.

Anergy is not a feature of in vitro and in vivo exhausted CTL

The high purity of T cells in our culture system raises the question whether the absence of CD28 costimulation induces anergy and this explains some of the phenotypes we observe. However, this is unlikely as our cells are cultured in the presence of IL-7 that has been suggested to prevent anergy and furthermore we still see exhaustion when cells are cultured with IL-2 (Fig 10A and 10B) another anergy preventing cytokine. Some CD28 costimulation, however, may be present in our culture system, as repeat peptide stimulated cells express CD80 mRNA levels 74-fold and 13-fold higher compared to unstimulated and single peptide stimulated cultures, respectively. Flow cytometry confirmed the high surface expression of CD80 on repeat peptide stimulated cells (Fig. 10C Fig).

To further exclude that our repeat peptide stimulated cells have features of anergy, we performed GSEA for sets of genes identified by in vitro anergy induction (gene set GSE5960). We find that both genes that are upregulated and downregulated in anergic T cells, are enriched in our repeat peptide stimulated cells (Fig 11A and 11B). Thus there is no evidence for the presence of anergy. We found a similar enrichment for genes both upregulated and downregulated in anergic T cells in in vivo exhausted gp33-specific CTL from LCMV clone 13 infected mice (gene set GSE41867) (Fig 11C and 11D).

Finally, we performed pathway enrichment analysis by IPA on the genes differentially expressed between repeat peptide stimulated cells and the unstimulated cells. We found that anergy related signaling pathway was neither up- or down-regulated in the repeat peptide stimulated cells (Fig 11E). In this comparison, the T cell exhaustion signaling pathway was again the most significant upregulated pathway. Taken together, these findings argue that anergy is not a major feature of our in vitro exhausted CTL.