Telomerase (TERT) is overexpressed in 80% to 90% of primary tumors and contributes to sustaining the transformed phenotype. The identification of several TERT epitopes in tumor cells has elevated the status of TERT as a potential universal target for selective and broad adoptive immunotherapy. TERT-specific cytotoxic T lymphocytes (CTL) have been detected in the peripheral blood of B-cell chronic lymphocytic leukemia (B-CLL) patients, but display low functional avidity, which limits their clinical utility in adoptive cell transfer approaches. To overcome this key obstacle hindering effective immunotherapy, we isolated an HLA-A2–restricted T-cell receptor (TCR) with high avidity for human TERT from vaccinated HLA-A*0201 transgenic mice. Using several relevant humanized mouse models, we demonstrate that TCR-transduced T cells were able to control human B-CLL progression in vivo and limited tumor growth in several human, solid transplantable cancers. TERT-based adoptive immunotherapy selectively eliminated tumor cells, failed to trigger a self–MHC-restricted fratricide of T cells, and was associated with toxicity against mature granulocytes, but not toward human hematopoietic progenitors in humanized immune reconstituted mice. These data support the feasibility of TERT-based adoptive immunotherapy in clinical oncology, highlighting, for the first time, the possibility of utilizing a high-avidity TCR specific for human TERT. Cancer Res; 76(9); 2540–51. ©2016 AACR.
The development of adoptive cell therapy (ACT) represents an emerging and realistic approach to treat cancer patients. This is testified by the numerous phase II clinical trials, the approval of specific T-cell therapies by the FDA, and the growing interest of biotechnology and pharmaceutical industry to generate “off-the-shelf” reagents to treat a large spectrum of tumors (1). At present, three types of ACT protocols can be defined based on isolated tumor-infiltrating T cells (TIL), T-cell receptor (TCR) engineered T cells, or chimeric antigen receptor (CAR) transduced lymphocytes, all of them already tested in clinical settings (2). TIL-based ACT can result in a long-lasting and complete cancer regression in metastatic melanoma patients (3, 4). However, this approach remains a personalized treatment that displays several technical constraints (5). The clinical response following adoptive TIL transfer was associated with T cells reactive toward mutated epitopes that were able to persist in patients for at least 1 month after lymphocyte infusion (6). These boundaries intrinsic to TIL-based ACT could be surmounted by gene therapy strategies based on genetically engineered lymphocytes where the desired TCR sequence insertion, by a virus-mediated delivery into naïve T cells, can confer an antigen-oriented immune specificity (7–9). To develop rapidly and apply ACT to a wide range of human neoplastic diseases, the characterization of high-avidity TCRs that efficiently and broadly recognize cancer cells is thus a primary goal (10). However, antitumor CTLs with a high-avidity TCR against non-mutated tumor-associated antigens (TAA) are normally deleted during thymus education of self-reactive T cells (11), and isolation of TCR recognizing individual mutations of patients' cancers is feasible in theory (12) but currently not applicable to large scale, standardized therapy. Nowadays, high-avidity TCR sequences could be achieved by different approaches. T cells with higher functional avidity could be generated in vitro by stimulation with autologous dendritic cells (DC) transfected with RNA encoding an allogeneic major histocompatibility complex (MHC) and the desired TAA (13). Alternatively, TCR can be isolated from mouse CTLs primed in vivo by vaccination of transgenic mice bearing human HLA-A2 molecules (14, 15), an approach that was recently improved by the immunization of human antigen–negative mice engineered to bear the whole human TCR-α and β gene loci together with the HLA-A2 allele (16). We previously reported the feasibility to isolate and enrich a polyclonal T-cell population specific for human telomerase (hTERT)865–873 epitope through in vitro stimulation of mouse T lymphocytes isolated from HLA-A2.1 transgenic mice (17). These CTLs recognized different hTERT-expressing human cancer cell lines, as well as colon cancer stem cells (17). Telomerase is reactivated in the majority of human tumors independently of their histology (18), and several hTERT epitopes, which are naturally processed and presented in association with MHC molecules on tumor cell surface, have been already documented (19–22). It is thus not surprising that TERT was ranked among the most prioritized TAAs (23), and several active immunotherapeutic approaches based on TERT antigen have been exploited to target, both in vivo and in vitro, either autologous or allogeneic antigen-presenting cells (APC), including antigenic peptides (24–27), RNA-based vaccines (28), as well as plasmid or viral vectors encoding hTERT (29). Unfortunately, clinical responses in these trials were limited, suggesting the need for more powerful, immune-based strategies. In the current study, we show the feasibility to transduce human T cells with a high-avidity mouse TCR able to recognize hTERT865–873 peptide in association with HLA-A2 molecules to control human solid tumors and hematologic malignancies, such as chronic lymphocytic leukemia (B-CLL). The high levels of hTERT in leukemic B cells correlated with poor clinic outcome (30, 31). We show here that hTERT-based ACT can selectively eliminate leukemic B cells, causing a minor toxicity against normal myeloid cells, making this approach suitable to clinical translation.
Materials and Methods
C57BL/6 (C57BL/6NCrl) mice were purchased from Charles River Laboratories Inc.; OT-1 (C57Bl/6-Tg(TcraTcrb)1100Mjb/J) and CD45.1+ mice (B6.SJL-PtrcaPepcb/BoyJ) from The Jackson Laboratory; NOG (NOD.Cg-Prkdcscid Il2rgtm1Sug/JicTac) and Rag2−/−/γc−/− mice (B10;B6-Rag2tm1Fwa II2rgtm1Wjl) from Taconic. The transgenic mice IgH.TEμ have been described previously (32), and their natural history of malignant progression has been monitored (Supplementary Fig. S1) in about 150 mice. All animal experiments were approved by the Verona University Ethical Committee, authorized by Ministerial Decree (16/2014-B) and conducted according to the guidelines of the Federation of European Laboratory Animal Science Associations.
Cell lines and patient samples
All cell lines were obtained from ATCC and were passaged for fewer than 6 months after their purchase. Human cell line identities were verified using short tandem repeat profiling. Healthy donor (HD) CD34+ cells derived from bone marrow (BM) were purchased from Lonza. Peripheral blood mononuclear cells (PBMC) isolated from B-cell chronic lymphocytic leukemia (B-CLL) patients and HDs were collected at the Hematology Unit, Azienda Ospedaliera Universitaria Integrata (AOUI) in Verona (Italy). All participating persons provided written informed consent in compliance with the Declaration of Helsinki. The study was approved by the local ethics committee (AOUI of Verona, n. 1496). Selected B-CLL patients had heterogeneous Binet clinical stages (33).
Generation of hTERT865–873– and hHCV1406–1415–specific, TCR-transduced T cells
OKT-3–activated PBMCs were infected with the viral supernatant of the hTERT865–873/PG13 cell line in the presence of hIL15 (100 μg/mL) and rIL2 (300 IU/mL; ref. 34). PBMCs were then immunomagnetically enriched for CD34 (Miltenyi) and expanded. Control HCV1406–1415 (KLVALGINAV)–specific, TCR-transduced T cells were generated following the same protocol. The percentage of CD4+ and CD8+ T cells (usually 20% and 80%, respectively) was always tested before in vitro or in vivo studies. In general, the amount of T cells used for in vivo treatments were adjusted in order to inject 2.5 × 106 CD8+ cells.
Generation and expansion of telomerase-specific T cells from B-CLL and HD PBMCs
B-CLL patients and HD were selected for HLA-A2 status, as assessed by FACS. T cells were immunomagnetically isolated from PBMCs. Human DCs were generated from CD14-selected monocytes (Miltenyi Biotec) and, after 100 μg/mL LPS maturation, pulsed with 10 μg/mL of the specific peptide. DCs were then used to stimulate T cells at an E/T ratio of 10:1 in complete RPMI-1640 in presence of IL7 (10 ng/mL), IL15 (2 ng/mL), and IL2 (10 IU/mL; all from Miltenyi Biotec). At days 7 and 15 of culture, T cells were restimulated with peptide-pulsed DCs. CD8+ T cells were screened for hTERT865–873 dextramer positivity and hTERT865–873–specific reactivity in IFNγ ELISA.
Systemic treatment of mouse leukemic chimeras
Chimeras were generated combining 106 CD45.2+ IgH.TEμ and 4 × 106 CD45.1+ syngeneic WT BM cells in Rag2−/−γc−/− mice, after preconditioning with Busulfan (25 mg/kg). When CD45.2+ cells raised to 15% of total B cells, 5 × 106 mTERT198–205– or OVA257–264–specific CTLs were intravenously injected twice in mice after γ-irradiation, followed by recombinant IL2 administration (17).
Systemic treatment of human leukemic chimeras
NOG mice were γ-irradiated (1.20 Gy) and subsequently engrafted with 105 human HLA-A2−CD34+ cells via tail-vein injection, as previously reported (35). Mice were then injected with 108 freshly isolated B-CLL PBMCs into retro-orbital plexus. Treatments with 2.5 × 106 hTERT865–873– or hHCV1406–1415–specific, TCR-transduced T lymphocytes were given 3 times, weekly, starting 1 week after the engraftment of the pathology and were always followed by IL2 administration. At sacrifice, organs were collected and analyzed by flow cytometry and IHC.
Data were indicated as the mean±SD. The Student t test was used to determine statistically significant differences between two treatment groups, while the ANOVA test was used in case of multiple comparisons. Growth curves were analyzed with repeated-measures (RM) ANOVA. Survival analysis was performed using the Kaplan–Meier survival analysis (log-rank) method. All P values less than 0.05 were considered statistically significant.
mTERT198–205–specific CTLs control mouse B-CLL progression
After repeated ACTs with polyclonal mouse (m)TERT198–205–specific CTLs to treat prostate cancers, we did not detect major side effects toward normal cells and tissues with the exception of a transient and reversible B-cell depletion in lymphoid organs (17). We therefore verified the therapeutic efficacy of mTERT198–205–specific CTL administration to restrain the expansion of monoclonal B cells, using the IgH.TEμ mouse model in which the sporadic SV40 large T antigen (hereafter indicated as SV40+) expression in mature B cells generates a B-CLL–like neoplasia (32). We considered mice as a leukemic (defined henceforward as IgH.TEμ+) when the majority of CD19+ B cells expressed the IgMb allele (IgMb > 75%). Splenic B cells isolated from IgH.TEμ+ mice displayed higher levels of TERT protein (Fig. 1A, top), as reported for human B-CLL (31), together with an increased TERT enzymatic activity (Fig. 1A, bottom) in comparison with B lymphocytes purified from either IgH.TEμ− or WT mice. Only B cells purified from IgH.TEμ+ mice were efficiently recognized by polyclonal mTERT198–205 CTLs, both in a cytofluorimetric cytotoxic assay (summarized in Supplementary Fig. S2A) and IFNγ release assay (Fig. 1B). These findings demonstrate that leukemic B lymphocytes can naturally process the endogenous TERT198–205 peptide and present it in a MHC class I–restricted fashion.
The low incidence of leukemia did not allow us to verify in vivo the therapeutic efficacy of a TERT-based ACT in IgH.TEμ mice, so we engrafted BM cells isolated from an IgH.TEμ+ mouse into immunodeficient Rag-2−/−γc−/− mice, partially ablated with chemotherapy, to establish a standardized B-CLL–like pathology. In fact, all engrafted mice developed a B-CLL with the same features of the IgH.TEμ+ mouse donor and tumor-bearing mice had to be euthanized 24 days after BM cell transplant (results for 2 passages out of 9 total are shown). The transplants did not affect the leukemic phenotype: the histological structure of the spleen was almost completely replaced by a uniform infiltration with SV40+CD19+CD5+B220low/− cells, similar to what occurred in the donor IgH.TEμ+ mouse (Supplementary Fig. S3A). Overall, B cells isolated from tumor-bearing chimeric mice showed high levels of mTERT expression and activity (Supplementary Fig. S3B) and were recognized by mTERT198–205–specific CTLs (Supplementary Fig. S3C). Therefore, the pre-leukemic B cells in BM were able to give rise to a full B-CLL–like disease.
In order to assess the in vivo efficacy of mTERT198–205–specific CTLs in immune-competent mice, we generated mouse chimeras by transplanting a mixture of BM cells, comprising one-fifth of leukemic cells (CD45.2+) derived from IgH.TEμ+ mouse (as previously described) and four-fifths of normal congenic BM-cells (CD45.1+) isolated from WT mice, into immunodeficient Rag-2−/−γc−/− mice. When mice had about 15% of CD19+ circulating cells, they were treated with two repeated transfers of either mTERT198–205– or the control, OVA257–264–specific CTLs (Supplementary Fig. S3D). The leukemic mice treated with mTERT198–205–specific CTLs displayed a significant reduction in the total number of blood circulating CD19+ B lymphocytes compared with control mice (Fig. 2A). However, mTERT-based ACT did not affect the normal B-cell reconstitution, because the levels of CD19+ B-cells in treated chimeras engrafted with only normal (WT) BM cells were not significantly different compared with untreated mice (Fig. 2A). In fact, mTERT-based ACT selectively reduced the CD19+CD45.2+ leukemic cell expansion without influencing the normal CD19+CD45.1+ B-cell development (Fig. 2B). This therapeutic effect of mTERT-specific ACT was mirrored by the significant inverse correlation between the frequencies of blood circulating, mTERT198–205–specific CTLs with the percentage of circulating B-CLL cells (Fig. 2C). The mTERT-specific ACT promoted a significant contraction in splenic SV40+ cells compared with control ACT, which was associated with an increased number of normal B220+ cells (Fig. 2D). Flow cytometry analysis on BM and spleen confirmed that, even if the overall percentage of splenic CD19+ cells was not modified by mTERT-specific ACT, the relative ratio between malignant CD45.2+ cells and healthy CD45.1+ cells significantly changed (Fig. 2E). As expected from this anti-leukemic activity, mTERT-specific ACT significantly improved the leukemia-bearing mouse survival (Fig. 2F).
Generation and characterization of engineered hTERT865–873–specific T cells
To verify the presence of a TERT-specific immune response induced by B-CLL progression, we selected B-CLL patients, sharing HLA-A2 allele and clinical stage (according to the Binet staging system), with at least 4-year follow-up; as a control, we included HLA-A2+ HDs. We isolated CD3+ T cells from patients and HD PBMCs and cocultured them with either hTERT865–873– or hCMV495–503–peptide pulsed (as positive control of stimulation), in vitro differentiated, HLA-A2–matched DCs. After 3 weekly in vitro stimulations, we verified the presence of hTERT-specific CD8+ T cells, defined as hTERT865–873 dextramer–positive cells, in approximately 50% of the B-CLL patients but not in HDs, even though in almost all patients and HDs we detected the presence of hCMV-specific CD8+ T cells (Fig. 3A). Patients whose T cells displayed positivity for hTERT-specific dextramer staining were also able to release IFNγ in response to hTERT865–873 peptide (Fig. 3B), with a linear relationship between cytokine production and the number of antigen-specific cells in culture (Fig. 3B, inset). Then, we examined the association between the presence of endogenous anti-hTERT865–873 response and disease progression, measured as time from diagnosis to initial therapy (time to first treatment, TTFT). Comparison between TTFT curves showed a trend to a shorter TTFT in the “TERT nonresponder” group (Fig. 3C). These data clearly unveiled the previously unknown presence of endogenous T cells specific for a physiologically processed hTERT865–873 epitope in B-CLL patients.
We therefore explored the feasibility to redirect human T lymphocytes by genetic engineering based on de novo expression of a high-avidity, HLA-A2–restricted, hTERT-specific TCR. Briefly, we first isolated a mouse CTL clone expressing a TCR for human HLA-A2–restricted hTERT865–873 epitope from a polyclonal, in vitro stabilized T cell population obtained from HLA-A2 transgenic mice vaccinated toward hTERT (17). The sequences of the α and β chains of the TCR (Fig. 3D) were then cloned by Vα- and Vβ-specific primer panels, amplified by PCR and the products of the reaction were sequenced and inserted into a retroviral vector (34) able to transduce naïve T cells. After expansion of human T cells in vitro transduced with either anti-hTERT or anti-hHCV (as control) TCR-encoding retroviruses (36), we checked the expression of the mouse TCR Vβ chain (mVβ3), a component of the hTERT865–873–specific TCR, on T-cell surface. (Fig. 3E). We further confirmed the presence of the correct, antigen-specific TCR on transduced T cells by dextramer staining. (Fig. 3E). The engineered hTERT865–873–specific T cells displayed a much higher avidity TCR compared with the endogenous TCRs expressed by T lymphocytes isolated from B-CLL patients, confirming the expected, intrinsic low avidity of the endogenous T-cell repertoire for self-antigens (Fig. 3F). Finally, we verified that hTERT865–873–specific TCR-engineered T cells did not cause a catastrophic MHC-restricted T-cell fratricide, as recently shown for survivin-specific, TCR-engineered T lymphocytes (37). The hTERT865–873–specific, TCR-engineered T cells recognized HLA-A2+ PBMCs and activated T-lymphocytes (infected with retrovirally encoded anti-hHCV TCR) only when the target T cells were pulsed with the hTERT865–873 peptide; no recognition of PBMCs and activated T lymphocytes, either unpulsed or pulsed with the control hHCV1406–1415 peptide, could be detected (Fig. 3G). These data indicate that hTERT865–873-specific, TCR-engineered T lymphocytes cannot kill human proliferating T lymphocytes.
Engineered hTERT865–873–specific, TCR-engineered T cells efficiently restrain human solid tumor and B-CLL in vivo growth
Both CD4+ and CD8+ T lymphocytes engineered with hTERT865–873–specific TCR showed an effector memory phenotype (CD45RA−CD62L−CCR7−; Supplementary Fig. S4A) characterized by high expression of CD69/CD44/CD38/CD25 and HLA-DR markers and low expression of exhaustion markers (LAG3/PD-1/Tim-3) compared with the specific isotype controls (Supplementary Fig. S4C). Moreover, CD4+ T cells were polarized toward IFNγ-producing Th1 cells (Supplementary Fig. S4D). We selected several tumor cell lines of different histology, comprising hematologic malignancies such as myeloma (U266), B-CLL (JVM13 and MEC-1), and Burkitt lymphoma cells (DG-75), and solid tumors, such as melanoma (SK23MEL), breast carcinoma (MDA-MB-231), and colon carcinoma (SW480) cells. The selected tumor cell lines showed high levels of TERT activity and heterogeneous levels of both TERT protein and HLA-A2 membrane complex (Fig. 4A). The hTERT865–873–specific, TCR-engineered T cells were capable of recognizing in vitro HLA-A2+ B-cell tumor cell lines with the expected exception of HLA-A2− MEC-1 (Fig 4B). The broad antitumor therapeutic activity of hTERT865–873–specific, TCR-engineered T cells was confirmed in vivo in immunodeficient NOG mice engrafted s.c. with five different HLA-A2+ tumors. Both adoptively transferred hTERT865–873– and hHCV1406–1415–specific, TCR-transduced T cells comparably migrated to the tumor site (Fig. 4C), although only hTERT-specific ACT significantly controlled tumor growth and prolonged mice survival (Fig 4D and Supplementary Fig. S4E). These data clearly advocate the potential effectiveness of TERT-based ACT approach to treat different human cancers.
We treated leukemic-bearing mice, systemically engrafted with JVM13-Luc cells, for 3 consecutive weeks with either hTERT865–873– or hHCV1406–1415–specific, TCR-transduced T cells, weekly monitoring tumor cell expansion by bioluminescence imaging. The hTERT-specific ACT induced a significant decrease in radiance signals in leukemia-bearing mice compared with mice receiving hHCV-specific ACT (Fig. 5A). Thirty-five days after the first treatment, we assessed tumor spreading to different organs (Fig. 5B–C). Moreover, we performed IHC of human CD20+ cells in the spleen of ACT-treated mice (Fig. 5D). All these analyses showed that engineered hTERT865–873–specific, TCR-engineered T cells induced a significant reduction in human B-CLL accumulation.
To further explore the usefulness of this immunotherapeutic approach from a translational standpoint, we evaluated the hTERT presence in PBMCs isolated from B-CLL patients and age-matched HDs. PBMCs of B-CLL patients exhibited higher hTERT expression and enzymatic activity levels compared with HD PBMCs (Supplementary Fig. S5A). Moreover, B cells isolated from HLA-A2+ B-CLL patients were also recognized in vitro by hTERT865–873–specific, TCR-engineered T lymphocytes similarly to the JVM13 cell line, while HLA-A2− B-CLL PBMCs as well as HLA-A2+ HD PBMCs were not recognized (Fig. 6A). We finally tested the anti-leukemic activity of ACT against B-CLL patients' PBMCs in vivo. To this aim, immunodeficient NOG mice, previously humanized with allogeneic BM-derived CD34+ cells, were engrafted with either HLA-A2+ or HLA-A2− PBMCs from B-CLL patients previously classified as TERT responders (Fig. 3C) and treated with three weekly courses of ACT (Supplementary Fig. S5B). In mice inoculated with HLA-A2+ PBMCs from B-CLL patients, the total number of circulating hTERT865–873–specific, TCR-engineered T cells inversely correlated with the percentage of CD19+HLA-A2+ malignant B cells (Fig. 6B, left). On the contrary, ACT with hTERT865–873–specific, TCR-engineered T cells had no significant effects on the number of hCD19+ cells developed in mice engrafted with HLA-A2− B-CLL (Fig. 6B, right). One month after B-CLL engraftment, we evaluated malignant B-cell infiltration in both spleen and BM. The hTERT-specific ACT in HLA-A2+ B-CLL–bearing mice visibly controlled the leukemia progression (Fig. 6C). Conversely, hTERT-specific ACT did not control the neoplastic progression in HLA-A2− B-CLL–bearing mice (Fig. 6D), although we detected the same amount of human CD8+ T-cell infiltration, as measured by flow cytometry (data not shown).
Adoptive transfer of hTERT865–873–specific, TCR-engineered T cells induces the depletion of mature granulocytes but does not affect hematopoietic stem cells
Even though engineered T cell–based therapies have shown long-term efficacy and promising curative potential for the treatment of cancer, several “on-target, off-tumor” toxicities have been reported (38). We therefore investigated the toxicity of hTERT-based immunotherapeutic approach toward the hematopoietic compartment. Although human CD34+HLA-A2+ cells were dimly recognized by hTERT865–873–specific, TCR-engineered T cells upon in vitro coculture (Fig. 7A), they maintained their in vitro ability to proliferate and differentiate into colonies in a semisolid medium (Fig. 7B) and preserved their multipotency once injected into γ-irradiated, immunodeficient NOG mice (Fig. 7C). Moreover, we treated human immune reconstituted (HIR) mice, obtained by transplantation of human HLA-A2+ CD34+ cells into immunodeficient NOG mice, with two consecutive ACTs (Supplementary Fig. S5C). Seventeen weeks after CD34+ cell engraftment, the total number of human CD45+ cells in the spleen was not affected by hTERT-specific ACT (Fig. 7D–E), while they were significantly reduced in the BM compared with mice receiving control ACT (Fig. 7E). The relative distribution among human leukocytes was maintained in the spleen (Fig. 7F, left), as well as the percentage of both human granulocytes and monocytes among myeloid cells (CD45+CD33+SSChigh and CD45+CD33+SSClow, respectively; Fig. 7F, right). Moreover, we could not identify any alteration of relative distribution of myeloid and lymphoid cells in BM (Fig. 7G, left), except for a significant contraction in human granulocytes (Fig. 7G, right). Indeed, among granulocytic cell subsets, we observed a nearly complete deletion of the more mature CD45+CD11b+CD16+ human myeloid cells in mice treated with hTERT-specific ACT (Fig. 7H). To confirm a potential hTERT-specific, ACT-linked toxicity on myeloid mature cell subsets, we purified from human BM aspirate of HDs the three main granulocytic maturation cell fractions (myeloblasts, promyelocytes, and granulocytes; ref. 39). Only the more mature cell population, i.e., CD45+CD11b+CD16+, was significantly recognized in vitro by hTERT865–873–specific, TCR-engineered T cells (Fig. 7I), although no difference in telomerase enzyme activity levels was detected among the different cell fractions (data not shown). Thus, hTERT865–873–specific, TCR-engineered T cells ignore progenitors and specifically target only mature myeloid subsets, potentially limiting the induction of a prolonged adverse neutropenia, a common consequence of conventional chemotherapeutic treatments.
We show here the strength and safety of a broadly applicable immunotherapy protocol based on engineered T cells generated by transduction with a high-affinity TCR capable of recognizing the complex formed by hTERT865–873 peptide and HLA-A2 molecule. Telomerase can be regarded as a universal TAA: it contributes to sustain tumor cell survival (40) and prevent apoptosis elicited by antiproliferative agents (41). Furthermore, hTERT is immunogenic (27) and, indeed, five HLA-A2–restricted epitopes have already been identified: I540 (19), R865 (20), 572Ya and 988Ya (42), and R38a (22). The first identified peptide, I540, was also tested in different cancer vaccination clinical trials that showed a modest impact on tumor control (26, 27). To the best of our knowledge, we show here for the first time that a specific cohort of B-CLL patients, with a trend toward a less aggressive leukemia, presented a specific, endogenous response toward the R865 epitope that was not detectable in HDs (Fig. 3A–B). Our data diverge from results published about I540 epitope, which is able to promote a comparable immune response in both tumor patients and normal donors (43). The differences between I540 and R865 epitopes might be related to a different baseline of immune tolerance between the two epitopes. Moreover, the TCR affinity of the endogenous anti-hTERT865–872 T cells isolated from B-CLL patients is deeply low (Fig. 3F), which makes it unrealistic to isolate high-avidity T cells from patients to develop an ACT protocol. So far, only one study reported the redirected T cell–based adoptive immunotherapy targeting hTERT with a TCR isolated from human CTLs (21), generated from PBMCs of HDs in vitro stimulated with HLA-A24:02–restricted nonameric hTERT461–469 peptide (44). However, the overall functional avidity to hTERT461–469 peptide evaluated for transduced T cells was about 10−7 mol/L, which is modest compared with the TCR affinity of our engineered T cells, which displayed 50% of their maximum response at 10−14 mol/L peptide concentration (Fig. 3F). Transduced anti-hTERT T cells specifically target tumor cells, such as leukemic B cells both in vitro and in vivo, without affecting normal B-cell development. Indeed, anti-TERT T cells do not alter dramatically stem cell differentiation, suggesting a lower toxicity compared with common chemotherapies that normally promote severe myeloid precursor depletion in treated patients, often requiring the administration of support therapy to restore normal hematopoiesis. The ability of TERT targeting to control B-CLL progression is also confirmed by our data using IgH.TEμ mice; in fact, polyclonal anti-TERT CTLs transfer was able to significantly improve the leukemia-bearing mouse survival (Fig 2I). The ability of hTERT865–873–specific, TCR-engineered T cells to eliminate mature granulocytes hints at an extension of this therapy toward human acute myeloid leukemia (AML), that is characterized by a strong TERT activity directly correlating with poor outcome of the disease (45). Another potential target for hTERT-specific ACT is represented by B-cell acute lymphoblastic leukemia, in which TERT locus is recurrently targeted by somatic chromosomal translocations (46) and TERT expression is a marker associated with inferior clinical outcome (47). Finally, our data about the therapeutic ability of hTERT-specific ACT to efficiently control cancer progression of different solid tumors (Fig. 4B) suggest how the transduced anti-hTERT T cells could be a potential “off-the-shelf” reagent applicable to treat many oncologic diseases.
TERT is physiologically activated in a limited number of human normal cell populations (48). Unfortunately, experimental models that can predict potential toxic effects against these human cells are currently not available and some off-target activity can be completely unpredictable (49). However, to control off-target activity and mitigate excessive in vivo T-cell activation/expansion after systemic infusion, which might induce a lethal cytokine storm (50), we plan to develop an antidote based on the administration of antibodies specific for the mouse Vβ3 chain of the engineered TCR. Moreover, to limit direct toxicity, uncontrolled growth and malignant transformation of hTERT865–873–specific, TCR-engineered T cells, we plan to insert a suicide gene, such as the inducible caspase-9 gene (iCasp9), which can be triggered in the case of unfavorable events (51). The clinic impact of these new strategies was recently demonstrated to control the graft versus host disease (GVHD) symptoms in acute leukemia relapsed patients, after allogeneic stem-cell transplant (52). Importantly, we also excluded a fratricide effect of engineered T cells, as well as the elimination of activated endogenous T cells, which might limit the power of immunotherapy. Finally, for a clinical point of view, it could be feasible to isolate a fully humanized TCR against hTERT by immunization of antigen-negative humanized mice that can generate optimal affinity TCRs for T-cell therapy (16).
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: S. Sandri, C. Cavallini, M.T. Scupoli, S. Sartoris, V. Bronte, S. Ugel
Development of methodology: S. Sandri, G. Fracasso, R.W. Hendriks, S. Ugel
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Sandri, F. De Sanctis, F. Boschi, G. Ferrarini, C. Cavallini, M.T. Scupoli, M.I. Nishimura
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Sandri, A. Lamolinara, F. De Sanctis, F. Boschi, C. Cavallini, M.T. Scupoli, M. Iezzi, M.I. Nishimura, V. Bronte, S. Ugel
Writing, review, and/or revision of the manuscript: S. Sandri, M.I. Nishimura, V. Bronte, S. Ugel
Study supervision: S. Sartoris, V. Bronte
Other (performed research and provided essential new reagents): S. Bobisse, K. Moxley
Other (assisted with experimental design of in vivo imaging study): A. Sbarbati
Other (analyzed tissue sections by immunohistochemistry and interpreted the data): A. Lamolinara
The authors thank Elisa Zoratti, Mauro Giacca, Lorena Zentilin, Martina Tinelli, Loredana Ruggeri, Ornella Poffe, Rosalinda Trovato, Alessandra Fiore, and Cristina Anselmi for technical help.
This work was supported by grants from the Italian Ministry of Health, Italian Ministry of Education, Universities, and Research (FIRB cup: B31J11000420001), Italian Association for Cancer Research (AIRC; grants 6599, 12182, and 14103), National Cancer Institute (P01 CA154778 to M.I. Nishimura), and Dutch Cancer Society (R.W. Hendriks).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.