Truncated form of TGFβRII, but not its absence, induces memory CD8⁺ T cell expansion and lymphoproliferative disorder in mice

Mice

CD4-dnTGFβRII, CD11c-dnTGFβRII, and CD4Cre mice on the C57BL/6 background were previously described (20, 30, 31). Rag1^−/−, b2m^−/−, OTI, OTII, and P14 mice were obtained from Jackson Laboratory. C57BL/6 and C57BL/6-Ly5.2/Cr (CD45.1) congenic mice were obtained from National Cancer Institute. Il15^−/− mice were purchased from Taconic. Tgfbr2^f/f, Il7^−/− and GzmB-Cre mice were kindly provided by H. Moses (Vanderbilt University), Schering-Plough (Palo Alto, CA), and J. Jacobs (Emory University), respectively (32–34). All animal experimentation was conducted in accordance with institutional guidelines.

Adoptive transfer and LM-OVA infection

Naive CD44^loCD62L^hiCD8⁺ OTI-Rag1^−/− cells were purified from spleen and peripheral lymph nodes (LNs) by FACS sorting. Every co-transfer experiment consisted of 0.5~1.0 × 10⁵ sorted naive OTI-control-Rag1^−/−, OTI-dnTGFβRII-Rag1^−/−, OTI-CD4Cre-Tgfbr2^f/f-Rag1^−/−, or OTI-GzmBCre-Tgfbr2^f/f-Rag1^−/− cells each on a different congenic background mixed at a 1:1 ratio. One day after adoptive transfer, mice were infected intravenously (i.v.) with 1 × 10⁵ recombinant Listeria monocytogenes expressing OVA (LM-OVA), which was a generous gift from H. Shen (University of Pennsylvania). Memory OTI-WT-Rag1^−/− or OTI-dnTGFβRII-Rag1^−/− cells were isolated from LNs and spleen of LM-OVA infected mice based on staining with CD8α and the corresponding congenic markers. For recall experiments, mice were infected i.v. with 5 × 10⁵ of the LM-OVA.

Flow cytometry

Fluorescent dye-labeled Abs against CD8α, TCRβ, CD45.1, CD45.2, CD43 (1B11), CD62L, CD69, CD25 were purchased from BioLegend, CD127, KLRG1, and CD122 were from eBioscience, IFN-γ, IL-2, and TNF were from BD Biosciences, and GzmB was from Caltag Laboratories. For intracellular cytokine staining, splenocytes from infected mice were stimulated with SIINFEKL peptide in the presence of Golgi-stop (BD Biosciences) for 4 hr, followed by fixing and permeabilization according to the manufacturer’s instructions (BD Biosciences). For cell division assay, purified memory OTI cells were stained with CFSE (Molecular Probes), and cultured in the presence of anti-CD3, PMA/ionomycin, IL-2, or IL-15 for 2–4 days. The cells were analyzed on a LSRII or FACSCariber (Becton Dickinson) and data were analyzed with FlowJo software (Tree Star).

In vivo CTL assay

OTI cells of each genotype were FACS sorted based on their congenic markers, and mixed with CFSE-labeled target cells at 1:5 ratio. The mixture of effector and target cells were injected i.v. into naive C57BL/6 hosts. Spleens from the recipient mice were harvested 14–16 hr later, and the ratio of CFSE^hi and CFSE^lo cells was measured by FACS analysis, and in vivo CTL activity was determined as described (35).

Isolation of lymphocytes from non-lymphoid organs

Peripheral blood lymphocytes (PBLs) were isolated from tail or retro-orbital vein of infected mice. Liver and lung cell homogenates were digested for 1 hour at 37°C with digestion buffer [RPMI + 5% FBS + 100 U/ml DNase I (Sigma) + 0.2mg/ml Collagenase IV (Sigma)]. Liver homogenates were centrifuged at 300 rpm for 3 min to remove hepatocytes, and lung homogenates were run through a 70μm filter mesh. Non-hepatic supernatant and lung lymphocytes were centrifuged at 1500 rpm for 10 min. The cell pellet was resuspended in 1 ml of RPMI + 5% FBS and mixed with 4 ml of 27.5% OptiPrep (Axis-Shield). To make a gradient, 1ml of RPMI was layered on top of the cells, and centrifuged at 2000 rpm for 20 min. Lymphocytes were removed from the interface of the gradient.

Real-time RT-PCR analysis

Total RNA was extracted with Qiagen mRNAEasy kit (Qiagen) according to the manufacturer’s instructions. RNA was reverse transcribed with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). The cDNA was subsequently used for quantitative PCR on an ABI Prism instrument using commercial primer-probe sets.

Karyotype analysis

Memory OTI-dnTGFβRII cells were isolated from LNs and spleen of LM-OVA infected mice, and stimulated with PMA and ionomysin for 2 days. Twenty mitoses were subjected to G-banding analysis (the Laboratory of Genome Biology and Cytogenesis, Van Andel Institute).

Statistical analysis

Unpaired Student’s t test was used to calculate the statistical significance of data comparison. p value of < 0.05 was considered to be significant.

Conditional deletion of Tgfbr2 in OTI cells results in increased number of SLECs, but does not lead to increased accumulation of total effector cells during acute Listeria infection

Using OTI cells expressing dnTGFβRII driven by the Cd4 promoter, we have shown that TGF-β signaling in CD8⁺ T cells controls the number of SLECs during acute Listeria monocytogenes expressing OVA (LM-OVA) infection (24). In this model, OTI cells isolated from dnTGFβRII-Rag1^−/− mice cannot be adoptively transferred into wild-type (WT) C57BL/6 recipients due to immunological rejection (24, 36). This problem is overcome by using CD11c-dnTGFβRII mice as hosts, as they also express human dnTGFβRII under the control of CD11c promoter (30). As reported previously (24), effector OT-dnTGFβRII cells expanded 2–3 fold more than OTI-WT cells by day 7 and 14 post infection (p.i.), and contracted normally by day 21 p.i. (Fig. 1A–B). An increased percentage of SLECs (KLRG^hiCD127^lo) in OTI-dnTGFβRII cells was also observed during the effector phase of LM-OVA infection (Fig. 1C).

Because of partial blockade of TGF-β signaling and the problem of immunological rejection in dnTGFβRII cells (18, 20), we used the TGFβRII conditional KO system as an alternate method to block TGF-β signaling in CD8⁺ T cells to study the effect of TGF-β on formation and function of memory CD8⁺ T cells. To delete TGFβRII in CD8⁺ T cells, we crossed Tgfbr2^f/f mice to OTI-Rag1^−/− as well as to mice expressing Cre recombinase driven by the Cd4 promoter (CD4Cre-Tgfbr2^f/f), and we used their littermates (CD4Cre-Tgfbr2^f/+ or CD4Cre-negative Tgfbr2^f/f) as controls. We adoptively co-transferred OTI-control or OTI-CD4Cre-Tgfbr2^f/f cells into naive C57BL/6 mice, and infected them with LM-OVA. The percentage of transferred cells was measured over the course of infection. We expected that the phenotype of CD4Cre-Tgfbr2^f/f cells would be much stronger than that of dnTGFβRII cells. However, in contrast to OTI-dnTGFβRII cells (Fig. 1A–B), the expansion of effector OTI-CD4Cre- Tgfbr2^f/f cells was comparable to control OTI cells during effector phase of acute LM-OVA infection (Fig, 1D–E). By contrast, similar to what was observed in OTI-dnTGFβRII cells (Fig. 1C), the OTI-CD4Cre-Tgfbr2^f/f cells had a higher percentage of SLECs and lower percentage of MPECs (Fig. 1F). These data suggest that dnTGFβRII cells and conditional TGFβRII KO cells have a different phenotype with regard to total effector T cell expansion, but show similarly increased percent of SLECs.

Late memory OTI-dnTGFβRII cells show enhanced accumulation after acute Listeria infection

Using the dnTGFβRII model, we next wanted to study the role of TGF-β signaling in the formation and function of memory CD8⁺ T cells. OTI-WT cells were either transferred alone into C57BL/6 mice or co-transferred together with OTI-dnTGFβRII cells into CD11c-dnTGFβRII hosts. At early time points, days 25–101 p.i., the frequency of transferred OTI-WT or OTI-dnTGFβRII cells among PBLs ranged from 0.01 to 8.73% in C57BL/6 or CD11c-dnTGFβRII hosts (Fig. 2A). Interestingly, by about day 300 p.i., CD11c-dnTGFβRII hosts that still retained OTI cells showed a dramatically elevated frequency of OTI-dnTGFβRII cells (Fig. 2A and 2B). The time at which this expansion could first be detected varied among the animals and for each experiment, and ranged from day 180 to 250 p.i. (data not shown). OTI-WT cells transferred into the CD11c-dnTGFβRII hosts behaved similarly to those transferred into C57BL/6 hosts, and the percentage of OTI-WT cells in PBLs was only slightly increased in either host (Fig. 2A). However, we observed that the engraftment efficiency of transferred OTI-WT cells in CD11c-dnTGFβRII hosts was significantly lower when compared to their engraftment into C57BL/6 hosts at 28–40 days p.i. (Fig. S1A). Although the exact reason for this lower engraftment of the OTI-WT cells in the CD11c-dnTGFβRII hosts is currently unclear, we hypothesize that this may be due to enhanced innate immunity in CD11c-dnTGFβRII hosts (30, 37), and a potential NK-mediated rejection of the OTI-WT cells over time. This long-term rejection of engrafted OTI-WT memory cells in CD11c-dnTGFβRII hosts is different than the early rejection of the OT-dnTGFβRII cells in C57BL/6 hosts, which occurs within days as previously reported (24, 36) (Fig. S1B–C). Since the long-term engraftment of OTI-WT cells was considerably reduced in CD11c-dnTGFβRII hosts, we had to use C57BL/6 hosts that received OTI-WT cells as our age-matched controls in most of the subsequent long-term experiments.

To measure the number of late memory OTI cells in different tissues, recipients were sacrificed at day 374–448 p.i. The CD11c-dnTGFβRII hosts that had received OTI-dnTGFβRII cells showed splenomegaly and lymphoadenopathy (Fig. 2C). The percentage of OTI-dnTGFβRII cells was significantly higher than that of OTI-WT cells in both lymphoid and non-lymphoid tissues (Fig. 2D), and greater than 20-fold more OTI-dnTGFβRII cells was recovered from each organ than OTI-WT cells (Fig. 2E). Unlike the transferred OTI-dnTGFβRII cells, the frequency of host CD8⁺ T cells from the spleen of CD11c-dnTGFβRII host did not increase over time (day 374–448 p.i.) (Fig. 2F), even though they had the CD44^hiCD62L^lo activated phenotype (Fig. 2G), suggesting that the OTI-dnTGFβRII phenotype in the CD11c-dnTGFβRII hosts is unique to the donor OTI cells that express the dnTGFβRII transgene.

Analyses of cell surface markers revealed that OTI-dnTGFβRII cells were phenotypically identical to memory OTI-WT cells before they exhibited lymphoproliferation (data not shown). However, although the expression of CD25, CD122 (IL-2Rβ/IL-15Rβ) and CD127, KLRG1, CD27, CD69, and CD44 was not different, the expanded memory OTI-dnTGFβRII cells showed reduced CD43, CD8α, and CD62L expression when compared with OTI-WT cells at about 300 days p.i. (Fig. 2H). As previously reported (29), these results demonstrate that over-expression of dnTGFβRII results in accumulation of dysregulated CD8⁺ T cells that phenotypically resemble memory CD8⁺ cells.

Deletion of Tgfbr2 does not lead to dysregulated expansion of late memory OTI cells upon Listeria infection

To eliminate any potential effect of TGF-β during T cell development and priming and to further focus on the role of TGF-β signaling on activated and memory CD8⁺ T cells, we crossed OTI-Tgfbr2^f/f-Rag1^−/− mice to GzmB-Cre animals (OTI-GzmBCre-Tgfbr2^f/f). Interestingly, although the OTI-GzmBCre-Tgfbr2^f/f cells had an increased percentage of SLECs as was seen in OTI-dnTGFβRII and OTI-Cd4Cre-Tgfbr2^f/f cells during effector phase of Listeria infection (Fig. 1C and 1F, and data not shown), the frequency of OTI-control and OTI-GzmBCre-Tgfbr2^f/f cells in PBLs was similar at day 7, 14 and 280 p.i. (Fig. 2I). At this late memory time point, almost all of the blood OTI-control cells were KLRG1 negative as expected; however, OTI-GzmBCre-Tgfbr2^f/f cells still showed a SLEC and double positive phenotype, as ~35% of the cells were still KLRG1⁺ (Fig. 2J), again confirming the role of TGF-β signaling in controlling the survival of the SLEC population. Furthermore, no significant difference was observed in the percentage of OTI-control and OTI-GzmBCre-Tgfbr2^f/f cells in the spleen even at day 438 p.i. (Fig. 2K). These results demonstrate that the accumulation of late memory OTI cells is unique to the dnTGFβRII model, and it is not recapitulated in the conditional KO system.

Development and proliferation of late memory OTI-dnTGFβRII cells is not dependent on IL-7 or IL-15

Given that IL-7 and IL-15 are important cytokines for memory CD8⁺ T cell homeostasis, we next examined whether dysregulated expansion of OTI-dnTGFβRII cells is dependent on IL-7 or IL-15 signaling. We previously reported that survival of effector OTI-dnTGFβRII cells is largely dependent on IL-15 signaling (24). However, we found that a small fraction of effector OTI-dnTGFβRII cells could differentiate into memory cells in Il15^+/+ or Il15^−/− CD11c-dnTGFbRII hosts, and they eventually massively expanded at 225 days p.i. (Fig. 3A and Fig. 3B). OTI-WT cells were hardly detected in the CD11c-dnTGFbRII co-transfer setting (Fig. 3A–3B), possibly as a result of out-competition of OTI-dnTGFβRII cells and reduced engraftment of OTI-WT cells (Fig. S1D). To further examine the role of IL-15 in OTI-dnTGFβRII cell expansion, we isolated these dysregulated memory OTI-dnTGFβRII cells 370 days post LM-OVA immunization, labeled them with CFSE, and re-transferred them into Il15^+/+ CD11c-dnTGFbRII or Il15^−/− CD11c-dnTGFbRII hosts. The proliferation of donor cells was assessed by cell division-induced loss of CFSE fluorescence. Memory OTI-dnTGFβRII cells proliferated similarly in IL-15 deficient and IL-15 sufficient hosts over a two-week period (Fig. 3C). These results suggest that in contrast to previous reports, IL-15 is not required for the development, survival, and dysregulated proliferation of late memory OTI-dnTGFβRII cells (29).

To determine whether IL-7 plays a critical role in memory OTI-dnTGFβRII cell dysregulated expansion, we co-transferred OTI-WT and OTI-dnTGFβRII cells into Il7^+/+ CD11c-dnTGFβRII or Il7^−/− CD11c-dnTGFβRII hosts. Interestingly, effector OTI-dnTGFβRII cells expanded at a much higher magnitude compared to OTI-WT cells in the IL-7 deficient environment than in the IL-7 sufficient environment (Fig. 4A). An increased percentage of SLECs in OTI-dnTGFβRII cells was also observed in IL-7 deficient hosts at day 7 and 12 p.i. (Fig. 4B). Furthermore, similar to the results with Il15^−/− CD11c-dnTGFβRII hosts (Fig. 3A), effector OTI-dnTGFβRII cells differentiated into memory T cells and eventually exhibited dysregulated expansion in Il7^−/− CD11c-dnTGFβRII mice, and this expansion occurred sooner than in Il7^+/+ hosts (Fig. 4C and 4D; compare day 146 p.i.). In addition, when CFSE-labeled dysregulated memory OTI-dnTGFβRII cells were transferred into Il7^+/+ CD11c-dnTGFβRII or Il7^−/− CD11c-dnTGFβRII host as performed in Fig. 3C, the proliferation of donor cells was accelerated in IL-7 deficient hosts, most likely due to the lymphopenic environment of the Il7^−/− hosts (Fig. 4E) (32). Overall, these results suggest that IL-7 is also not essential for the development and dysregulated proliferation of late memory OTI-dnTGFβRII cells.

Dysregulated memory OTI-dnTGFβRII cells can proliferate independent of MHC class I

Memory CD8⁺ T cells can persist in the absence of antigen (1, 3, 4). To examine whether proliferation of dysregulated memory OTI-dnTGFβRII cells is dependent on MHC class I, we isolated massively accumulated memory OTI-dnTGFβRII cells from infected CD11c-dnTGFβRII hosts as performed in Fig. 3C, and transferred (2 × 10⁶ cells) into sublethally irradiated C57BL/6 or b2m^−/− mice, which are deficient in the expression of functional MHC class I. At day 30 after transfer, the frequency of transferred OTI-dnTGFβRII cells in both spleen and LNs was greater in b2m^−/− mice most likely due to the lymphopenia caused by the absence of host CD8⁺ T cells (Fig. 5A and 5B). In addition, dysregulated memory OTI-dnTGFβRII cells proliferated extensively in both irradiated C57BL/6 and b2m^−/− mice (Fig. 5C), indicating that dysregulated memory OTI-dnTGFβRII do not need MHC class I to survive and proliferate.

OTI-dnTGFβRII cells, but not OTI-CD4Cre-Tgfbr2^f/f cells, show dysregulated expansion under steady state condition

C57BL/6 mice expressing dnTGFβRII driven by the Cd2 promoter developed polyclonal memory-like CD8⁺ T cell hyperproliferation under steady state condition, without developing autoimmunity (29). However, our dnTGFβRII mice on the Rag1^+/+ background did not show this phenotype as they develop colitis so rapidly it was not possible to assess this (20). However, we did observe the hyperproliferation phenotype in pathogen-induced memory OTI-dnTGFβRII-Rag1^−/− cells (Fig. 2). To eliminate the autoimmune phenotype in dnTGFβRII mice as well as the complications associated with the CD11c-dnTGFβRII hosts, we analysed the phenotype of OTI-dnTGFβRII cells on the Rag1^−/− background. The expression of CD44 on OTI cells was analyzed in OTI-dnTGFβRII-Rag1^−/− and OTI-CD4Cre-Tgfbr2^f/f-Rag1^−/− mice under steady state conditions, in which endogenous T cell receptors are excluded and autoimmunity was eliminated. Despite the fact that OTI-dnTGFβRII-Rag1^−/− mice do not express the cognate OVA antigen, a large proportion (35~95%) of OTI-dnTGFβRII cells isolated from LNs and spleen were CD44^hi memory phenotype by 13 weeks of age, whereas OTI-WT cells retained the CD44^lo naive phenotype even at 32 weeks of age (Fig. 6A and data not shown). In contrast to OTI-dnTGFβRII cells, more than half of OTI-CD4Cre-Tgfbr2^f/f cells, similar to control cells, still retained CD44^lo naive phenotype at 11–25 weeks of age (Fig. 6B). Interestingly, OTI-dnTGFβRII-Rag1^−/− mice displayed a considerable increase in the absolute number of OTI cells in both LNs and spleen by 12–32 weeks of age (Fig. 6C). Most of the expanded OTI cells were CD44^hi memory-like cells but CD44^lo naive phenotype cells were also expanded in some of the animals (Fig. 6A and 6C). Furthermore, the dysregulated dnTGFβRII cell expansion is not OTI-Tg line specific, as it was also seen in lymphocytic choriomeningitis virus-specific CD8⁺ P14-dnTGFβRII cells (Fig. S2A–D). As expected, consistent with the late memory OTI cell phenotype, this dysregulated expansion of OTI cells was detected only when TGF-β signaling was blocked by over-expression of dnTGFβRII, and not when the whole receptor was deleted (Fig. 6D). Indeed, no significant difference in the absolute number of OTI cells in the LNs and spleen was observed between control and OTI-CD4Cre-Tgfbr2^f/f-Rag1^−/− mice (Fig. 6D), further demonstrating that the deletion of the TGFβRII itself, at the same time as when the dnTGFβRII would be expressed (thymic double positive stage) does not lead to dysregulated expansion of memory-like OTI cells, even though a fraction of naive OTI-CD4Cre-Tgfbr2^f/f cells do become CD44^hi over time (Fig. 6B).

Rapid differentiation of naive OTI-dnTGFβRII cells into CD44^hi memory-like cells occurs even in the absence of endogenous Tgfbr2 gene

Because the effects of TGF-β signaling in T cell differentiation is context dependent (38), it is possible that a low level of leaked TGF-β signaling in OTI-dnTGFβRII cells promotes CD44^hi memory CD8⁺ T cell differentiation, which then somehow causes progressive expansion. To examine this possibility, we crossed OTI-dnTGFβRII-Rag1^−/− mice with CD4Cre-Tgfbr2^f/f-Rag1^−/− mice to completely delete TGF-β signaling in naive OTI cells. We still found an increased percentage of CD44^hiCD122^hi memory-like cells in the LNs of OTI-dnTGFβRII-Rag1^−/− × CD4Cre-Tgfbr2^f/f-Rag1^−/− mice when compared with OTI-CD4Cre-Tgfbr2^f/f cells (Fig. 6E), suggesting a dominant effect of truncated form of TGFβRII on CD8⁺ T cell expansion that is independent of blocking TGF-β signaling in these cells.

Memory OTI-dnTGFβRII cells develop lymphoma

As memory-like CD8⁺ T cells that express dnTGFβRII under the control of the Cd2 promoter have been shown to develop lymphoma (28), we analyzed whether over-expression of dnTGFβRII by Cd4 promoter also causes cellular transformation. Dysregulated memory-like OTI-dnTGFβRII cells or Listeria-induced memory OTI-dnTGFβRII cells were isolated from 6 months old OTI-dnTGFβRII-Rag1^−/− mice or CD11c-dnTGFβRII hosts at day 420 p.i., respectively, and karyotype analysis was performed. These cells showed chromosomal abnormality, including trisomy of chromosome 15 and reciprocal translocation between chromosome 4 and 15. (Table 1 and Fig. 7A). Gain of chromosome 15 is frequently present in several mouse models of lymphoma (39–41), and it has been suggested that Myc protooncogene located on mouse chromosome 15 contribute to trisomy 15 and leukemic transformation (42). We therefore analysed Myc expression in memory-like OTI-dnTGFβRII cells from 2–3 months old OTI-dnTGFβRII-Rag1^−/− mice before they exhibited dysregulated expansion. We found that although the expression level of Myc target genes, such as Id2 (inhibitor of DNA binding 2) (43) and Cdkn1a (cyclin-dependent kinase inhibitor 1A; also called p21) (44) were increased in CD44^hi memory-like OTI-dnTGFβRII cells, Myc expression did not differ between these groups (Fig. 7B). However, when we analyzed dysregulated memory-like OTI-dnTGFβRII cells isolated from OTI-dnTGFβRII-Rag1^−/− mice at 6 months of age, Myc expression was increased in memory-like OTI-dnTGFβRII cells compared with that of OT-I WT cells (Fig. 7C). These results suggest that over-expression of dnTGFβRII in OTI cells by Cd4 promoter results in cellular transformation most likely through a Myc-mediated mechanism.

Transformed memory OTI-dnTGFβRII cells produce effector cytokines and have CTL activity, but can not proliferate after recall response

To determine whether transformed memory OTI-dnTGFβRII cells retain their functionality as memory cells, we assessed the capacity of these cells to produce effector cytokines, and to kill antigen-sensitized target cells in vivo. We isolated memory OTI-WT cells or transformed memory OTI-dnTGFβRII cells from infected C57BL/6 or CD11c-dnTGFbRII mice at day 380 p.i,, respectively, and re-stimulated them with SIINFEKL peptide ex vivo. Interestingly, the transformed memory OTI-dnTGFβRII cells were functional in terms of IFN-γ, TNF and GzmB production, when compared to memory OTI-WT cells (Fig 8A). Furthermore, when the same number of each memory OTI population was re-transferred into naive hosts along with SIINFEKL-pulsed target cells, no difference was observed in the in vivo ability of OTI-dnTGFβRII cells to kill target cells when compared to memory OTI-WT cells (Fig. 8B), suggesting that they also retain their cytolytic activity. Similar results were also obtained when we used early memory OTI-WT and OTI-dnTGFβRII cells that had not yet exhibited dysregulated expansion (Fig. S3A and S3B). Similarly, early memory OTI-CD4Cre-Tgfbr2^f/f cells exhibited a normal ability to produce IL-2 and IFN-γ after SIINFEKL peptide stimulation ex vivo (Fig. S4A). Interestingly, however, transformed memory OTI-dnTGFβRII cells proliferated less than memory OTI-WT cells in response to SIINFEKL peptide or anti-CD3 ex vivo, while they responded similar to OTI-WT cells to PMA/ionomycin stimulation (Fig. 8C). Furthermore, transformed memory OTI-dnTGFβRII cells did not respond to IL-2 or IL-15 stimulation (Fig. 8D), whereas early memory OTI-dnTGFβRII cells isolated from the hosts at day 70 p.i. proliferated similarly to OT-WT cells in response to these cytokines (Fig. S3C). To determine the ability of transformed memory OTI-dnTGFβRII cells to respond to a secondary infection in vivo, we isolated transformed memory OTI-dnTGFβRII cells and memory OTI-WT cells at day 461 p.i., and re-transferred the same number of cells from each into naive hosts followed by LM-OVA infection. Consistent with results from ex vivo experiments, OTI-dnTGFβRII cells responded poorly to a secondary infection, whereas memory OTI-WT cells mounted a proliferative response (Fig. 8E). In contrast, no difference in proliferative response was observed when we transferred memory OTI-WT and OTI-dnTGFβRII cells isolated 180 days p.i. that had not exhibited dysregulated expansion (Fig. S3D). The early memory OTI-control and OTI-GzmBCre-Tgfbr2^f/f cells also expanded similarly in vivo upon re-challenge (Fig. S4B). These results indicate that blocking TGF-β signaling on memory CD8⁺ T cells does not impact their function and recall response. In addition, although transformed late memory OTI-dnTGFβRII cells are capable of producing effector cytokines and killing target cells, they have defects in antigen- or cytokine-induced proliferative responses. The consequence of long-term over expression of dnTGFβRII should be significantly considered for studies where the over expression of this construct is used for enhancing T cell function during T cell mediated immunotherapy.