Acute myeloid leukemia (AML) bears heterogeneous cells that can consequently offset killing by single-CAR-based therapy, which results in disease relapse. Leukemic stem cells (LSCs) associated with CD123 expression comprise a rare population that also plays an important role in disease progression and relapse. Here, we report on the robust anti-tumor activity of a compound CAR (cCAR) T-cell possessing discrete scFv domains targeting two different AML antigens, CD123, and CD33, simultaneously. We determined that the resulting cCAR T-cells possessed consistent, potent, and directed cytotoxicity against each target antigen population. Using four leukemia mouse models, we found superior in vivo survival after cCAR treatment. We also designed an alemtuzumab safety-switch that allowed for rapid cCAR therapy termination in vivo. These ﬁndings indicate that targeting both CD123 and CD33 on AML cells may be an effective strategy for eliminating both AML bulk disease and LSCs, and potentially prevent relapse due to antigen escape or LSC persistence. Introduction AML is a hematological disease characterized by the malignant transformation and hyperproliferation of imma- ture myeloid cells, which replace normal bone marrow cells. These authors contributed equally: Jessica C. Petrov and Masayuki Current chemotherapy regimens that combine cytarabines Wada. with anthracyclines successfully treat few patients and even Electronic supplementary material The online version of this article fewer with relapsed and/or refractory AML [1–3]. Allo- (https://doi.org/10.1038/s41375-018-0075-3) contains supplementary geneic hematopoietic stem cell transplantation (HSCT) material, which is available to authorized users. remains the only viable treatment option for AML, and only * Masayuki Wada a limited number of patients qualify . Moreover, 50–70% firstname.lastname@example.org of patients relapse after chemotherapy and HSCT, with the * Yupo Ma 5-year survival rate at a dismal 27%. Considering the email@example.com shortcomings of current AML therapy and the stagnation of treatment advances in the past few decades, new therapies iCell Gene Therapeutics LLC Research & Development Division are desperately needed. Long Island High Technology Incubator, 25 Health Science Drive, Stony Brook, NY 11790, USA CAR T-cell immunotherapy is a new and powerful therapy that has already shown utility as a curative treat- Department of Hematology West China Hospital, Sichuan University, Chengdu, P.R. China ment for malignant hematological diseases, most notably B- cell lymphomas and plasma cell malignancies through tar- Department of Pathology Stony Brook Medicine Stony Brook, NY 11794, USA geting CD19 and BCMA, respectively [5, 6]. However, substantial relapse is seen in patients one year after CAR Macau Institute for Applied Research in Medicine and Health, Macau University of Science and Technology, Macau, China therapy. Therefore, a single target for CAR-based treatment may not be sufﬁcient to prevent disease relapse. It follows Department of Internal Medicine Stony Brook Medicine, Stony Brook University Medical Center, Stony Brook, NY 11794, USA that compound targeting of more than one antigen repre- sents a critical need to improve CAR therapy outcomes. Department of Hematology, Chengdu Military General Hospital, Chengdu, Sichuan, P.R. China 1234567890();,: 1318 J. C. Petrov et al. Translating CAR T-cell therapy to AML also requires a and leukemia blasts in patient samples. As a safety-switch careful understanding of characteristics unique to the dis- to protect against the high potency of our cCAR, we ease, and the components which drive it. AML is char- developed a strategy that rapidly eliminated all residual acterized by the presence of heterogeneous blast cells, cCARs in leukemia mouse models. Together, this study which are highly aggressive rapidly dividing cells that form supports the development of 123b-33bcCAR as a pro- the bulk of disease. AML is uniquely challenging to treat mising immunotherapy for AML. due to the role of leukemic stem cells (LSCs) in initiating and maintaining disease . LSCs remain unaffected by chemotherapies targeting rapidly dividing cells due to their Materials And Methods quiescent nature. A successful CAR therapy for AML would target two separate antigens to both: (1) combine the Construction of xenogenic mouse models bulk targeting of heterogeneous malignant cells with elim- inating LSCs that cause relapse and (2) provide coverage of Four mice models were used to analyze anti-leukemia effect multiple targets to limit single-antigen relapse. and compound antigen targeting: (1) MOLM13 AML tumor CD33 is a myeloid marker that has been a target of great model N = 4, (2) U937 AML tumor model N = 4, (3) Jur- interest in the treatment of AML due to its speciﬁc katxp123 single-antigen model N = 2, and (4) Jurkatxp33 expression on bulk AML disease and minimal expression single-antigen model N = 2. Male 12-week-old NSG mice on normal cells [1, 8–10]. Patients treated with gentuzumab (NOD.Cg-Prkdcsid Il2rgtm1Wjl/SzJ) were purchased from ozogamicin, an anti-CD33 antibody therapy, relapsed with the Jackson Laboratory (Bar Harbor, ME) and used under a CD33 AML [8, 11]. Thus, while targeting CD33 elim- Stony Brook University IACUC-approved protocol. NSG inates the majority of disease, supplementing with an mice were given the standard preconditioning sublethal additional target would help eliminate CD33 leukemic (2.0 Gy) dose of irradiation and then intravenously injected cells or disease-replenishing LSCs. A study of 319 AML on Day 0 with (1) 1.0 × 10 MOLM13 cells per mouse, (2) 6 6 patients found that 87.8% of AMLs expressed CD33 . 1.0 × 10 U937 cells, (3) 1.0 × 10 Jurkatxp123 cells, or (4) CD123 is also widely present in AML blasts and the same 1.0 × 10 Jurkatxp33 cells per mouse. Three days (Day 3) 319 AML patient study found that 9.4% of AMLs express following tumor cell injection, mice were intravenously CD123 without CD33. Therefore, targeting CD33 and injected with one dose of 10 × 10 123b-33bcCAR T-cells CD123 together may prevent antigen escape associated with or control T-cells. relapse. Note: For all mice studies, one dose is deﬁned as total CD123 (alpha chain of the interleukin 3 receptor) is an cells injected; actual dose of cCAR T-cells is lower due to ideal target, as it is overexpressed in AML [12, 13]. efﬁciencies of gene transfer ~25%. Luciferin injection and + − Importantly, it displays high expression on CD34 CD38 IVIS imaging were accomplished as previously described LSCs and is absent from or minimally expressed on normal [17–20]. + − hematopoietic stem cells (HSCs) [14–16]. CD34 CD38 Additional methods (including in vitro assays and ﬂow cells are deﬁned as LSCs since they can initiate and cytometry methods) are described in supplementary mate- maintain the leukemic process in immunodeﬁcient mice. rials. The ﬂow cytometry antibodies used are listed in + − + The number of CD34 CD38 CD123 LSCs is predictive Supplementary Table 1. of treatment outcomes for AML patients . Although AML is a heterogeneous disease, the majority of AML samples express either CD33, CD123, or both [1, 13]. Results Targeting both CD123 and CD33 would, therefore, elim- inate AML in the majority of patients. Generation of CD123-CD33 cCAR (123b-33bcCAR) T- In our preclinical study, we designed a CD123b- cells CD33b cCAR expressing discrete anti-CD123 and anti- C33 CAR units to target bulk disease and LSCs simul- The 123b-33bcCAR T-cells were generated by transduction taneously in AML. Moreover, dual targeting offers more of primary peripheral blood T-cells with the lentiviral comprehensive ablation and may overcome the pitfalls of construct depicted in Fig. 1a. Flow cytometry analysis single-antigen therapy by preventing relapse due to showed that ~25% of T-cells expressed the CD123 or CD33 antigen loss. We showed that CD123b-CD33b cCAR CAR F(Ab′)2 fragment after transduction (Fig. 1b). This (123b-33bcCAR) T-cells speciﬁcally ablated leukemic 123b-33bcCAR was designed to deplete AML cells by + − + cells expressing either or both CD123 and CD33 in vitro targeting both bulky disease and the CD34 CD38 CD123 and in vivo. We also found that the 123b-33bcCAR leukemic population to eliminate active disease and prevent displayed remarkable efﬁcacy in eliminating both LSCs leukemia relapses (Fig. 1c). Anti-CD123 and CD33 compound CAR T-cells target acute myeloid 10 leukemia 1319 Fig. 1 123b-33bcCAR A. Construct and Expression. a Two discrete CAR units: an anti- CD123b CAR comprised of: a CD8-derived hinge (H) and transmembrane (TM) regions, and either 4-1BB or CD28 co- activation domains linked to the CD3ζ signaling domain is fused B. Non-transduced to a complete anti-CD33b CAR Control T-cells 123b-33bcCAR T-cells activated T-cells by a self-cleaving P2A peptide. A strong spleen focus forming virus promoter (SFFV) and a CD8 leader sequence were used for efﬁcient expression of the 123b-33bcCAR molecule on the T-cell surface. b Expression was measured by FACS against control T-cells. c 123b- 33bcCAR T-cell dual-pronged approach schematic CD3 C. 123b-33bcCAR T-cells effectively lyse acute myeloid Leukemic cell phenotypes are included in Supplementary leukemia cell lines Fig. S1. To assess the cytotoxicity of 123b-33bcCAR T-cells, we 123b-33bcCAR T-cells’ discrete receptor units conducted co-culture assays against AML cell lines independently lyse target cells in an antigen- + + − + MOLM13 (CD123 CD33 ) and U937 (CD123 CD33 ). speciﬁc manner FACS analysis of 123b-33bcCAR cytotoxicity in 24 h co- cultures revealed virtually complete lysis (>90%) of To further conﬁrm our cCAR’s independent antigen tar- MOLM13 and U937 tumor cells at E (effector):T(target) geting ability, T-ALL Jurkat cell lines were overexpressed ratios of 2:1 and 5:1 (Fig. 2a, Supplementary Fig. S4A). with either CD123 (Jurkatxp123) or CD33 (Jurkatxp33) We further evaluated the dose-dependent cytotoxicity of and independent antigen expression was conﬁrmed (Sup- the 123b-33bcCAR by progressively decreasing the E:T plementary Fig. S1). To determine targeting functionality, ratio against two other AML cell lines: KG1a (CD123 123b-33bcCAR T-cells were co-cultured against these + + dim + CD33 ) and HL60 (CD123 CD33 ). Using E:T ratios cells in addition to wild-type Jurkat cells (Fig. 2b). We of: 0.25:1, 0.5:1, 1:1, 2:1, 5:1, and 10:1, 123b-33bcCAR found that the 123b-33bcCAR T-cells speciﬁcally and displayed robust lysis even at the 0.25:1 ratio (Fig. 2c). potently ablated (>90% lysis) Jurkatxp123 and Jurkatxp33 F(Ab’)2 1320 J. C. Petrov et al. 123b-33bcCAR T-cells demonstrate targeted lysis of tumor cell lines Fig. 2 123b-33bcCAR T-cells demonstrate targeted lysis of A. Population depletion of MOLM13 tumor cells Population depletion of U937 tumor cells 2:1 5:1 2:1 5:1 tumor cells lines. All target populations are encircled. a Flow cytometry analysis of control T-cells and 123b- 33bcCAR T-cells against + + MOLM13 (CD123 CD33 )or U937 (CD123 CD33 + ) tumor target cells at 2:1 and 5:1 E:T ratios. b Flow cytometry analysis of control T-cells and 123b-33bcCAR T-cells against wild-type (WT) Jurkat tumor cells and Jurkat cells expressing CD33 CD33 either CD123 (Jurkatxp123) or CD33 (Jurkatxp33) at a 2:1 E:T B.Population depletion of Jurkat xp123 cells Population depletion of Jurkat xp33 cells ratio. c Dose-dependent cultures WT Jurkat Jurkat xp123 WT Jurkat Jurkat xp33 performed with HL60 dim + (CD123 CD33 ) and KG1a + + (CD123 CD33 ) cells display high cCAR killing efﬁciency at E:T ratios ranging from 0.25:1 to 10:1 CD123 CD33 C.Dose-dependent cocultures .25:1 1:1 2:1 5:1 10:1 .25:1 1:1 2:1 5:1 10:1 E:T ratios HL60 KG1a cells when compared to wild-type Jurkat cells expressing PT4:B-ALL) (Fig. 3, Supplementary Fig. S4C). Compared neither antigen (Supplementary Fig. S4B). Overall, 123b- to the previous anti-tumor cytotoxicity against AML cell 33bcCAR T-cells displayed effective bulk cytotoxicity, lines (Fig. 2), 123b-33bcCAR T-cells showed similarly ablating cell populations expressing varying combinations potent anti-leukemic activity against all patient samples, of targets, and cells expressing only each individual anti- with >80% tumor lysis at the 2:1 ratio and >98% tumor gen (Fig. 2). lysis at the 5:1 ratio. We also examined the ability of our 123b-33bcCAR to 123b-33bcCAR T-cells effectively lyse primary eliminate speciﬁc cell populations including LSCs + + − leukemia tumor cells (CD123 CD34 CD38 ) in the PT3 sample and myeloid variable + leukemia bulk disease (CD34 CD33 )inthe We next established the anti-tumor properties of the 123b- PT4 sample (Fig. 3c, d). We found that 123b-33bcCAR 33bcCAR against four primary patient leukemia samples, T-cells successfully ablated both LSCs and bulk disease + + + including two CD123 CD33 AML and two CD123 B- cells. Patient sample phenotypes are included in Sup- ALL samples (PT1:AML, PT2:B-ALL, PT3:AML, and plementary Fig. S2. CMTMR CD3 Experiment Control Experiment Control Percent Lysis CMTMR CD3 Experiment Control Control Experiment Anti-CD123 and CD33 compound CAR T-cells target acute myeloid 10 leukemia 1321 123b-33bcCAR T-cells demonstrate targeted lysis of primary patient tumor cells Fig. 3 123b-33bcCAR T-cells A. Population depletion of PT1 (AML) tumor cells B. Population depletion of PT2 (B-ALL) tumor cells demonstrate targeted lysis of 2:1 5:1 2:1 5:1 primary patient tumor cells. All target populations encircled. a Flow cytometry analysis of control T-cells and 123b- 33bcCAR T-cells against PT1 tumor target cells at 2:1 and 5:1 E:T ratios. b Flow cytometry analysis of control T-cells and 123b-33bcCAR T-cells against PT2 tumor target cells at 2:1 and 5:1 E:T ratios. c Flow cytometry analysis of control T-cells and CD33 CD123 123b-33bcCAR T-cells against PT3 tumor target cells at 2:1 and C. Population depletion of PT3 leukemic stem cells 5:1 E:T ratios. The target cell 2:1 5:1 2:1 5:1 + + population (CD123 CD34 )is further broken down by CD38 expression to display LSC + + - (CD123 CD34 CD38 ) elimination. d Flow cytometry analysis of control T-cells and 123b-33bcCAR T-cells against PT4 tumor target cells (CD33 bulk disease) at 2:1 and 5:1 E:T ratios CD34 CD34 D. Population depletion of PT4 bulk disease 2:1 5:1 CD34 123b-33bcCAR T-cells exhibit profound anti-tumor injected with 123b-33bcCAR T-cells had ~92-99% less activity in vivo tumor burden than control mice by Day 13. Moreover, cCAR-treated mice also had signiﬁcantly more favorable In order to evaluate the in vivo anti-tumor activity of 123b- survival outcomes (P-value = 0.0082 for Mantel–Cox in 33bcCAR T-cells, we developed two mouse models with both groups) (Fig. 4c, f). + + either luciferase-expressing MOLM13 (CD123 CD33 )or We also evaluated tumor cell and CAR T-cell persistency − + U937 (CD123 CD33 ) cells to induce ﬂuorescence visible at the time of sacriﬁce. Peripheral blood was collected and tumor formation (Fig. 4a, d). The 123b-33bcCAR cells analyzed via ﬂow cytometry for the presence of transplanted signiﬁcantly reduced tumor burden and prolonged survival tumor (MOLM13 or U937 cells) and T-cells (cCAR or in MOLM13 injected and U937-injected mice. Mice were control). While control-treated mice showed signiﬁcant given a single dose of 123b-33bcCAR T-cells or control T- residual tumor populations (~75–87%) in the peripheral cells cells and tumor burden assayed by IVIS imaging. blood, 123b-33bcCAR treated mice showed virtual deple- There was a signiﬁcant difference (P < 0.0001) in IVIS tion of all tumor cells (Supplementary Fig. S3). In addition, measurement of tumor burden between the control group 123b-33bcCAR treated mice showed signiﬁcant T-cell and the 123b-33bcCAR treatment group from Day 6 expansion and persistence, whereby virtually all the onwards in both xenograft models (Fig. 4b, e). Mice human cells found in the peripheral blood were cCAR T- CD123 CD3 Experiment Control Experiment Control CD33 Experiment Control CD38 CD3 Experiment Control Control Experiment 1322 J. C. Petrov et al. 123b-33bcCAR T-cells suppress MOLM and U937 tumor growth in vivo Fig. 4 123b-33bcCAR T-cells demonstrate a profound anti- A. IVIS imaging of MOLM13-Luc injected murine model MOLM13 tumor burden B. leukemic effect against MOLM13 and U937 cell lines in Control T-cells 123-33cCAR T-cells 1.5 x 10e5 two in vivo xenograft mouse models. a Mouse model using Day 3 MOLM13 cells to induce 1.5 x 10e4 ** * measurable tumor formation. 1.0 x 10e6 Mice treated with either 123b- 33bcCAR T-cells or control T- Day 6 Days after injection cells. Tumor burden visualized 1.0 x 10e5 on days 3, 6, 9, and 13 (left 5.0 x 10e6 C. Survival Proportions: MOLM13 mice panel) with b graphical Day 9 Control T-cells representation on the right. c cCAR T-cells 5.0 x 10e5 Log-rank Mantel–Cox test 5.0 x 10e7 shows signiﬁcance for improved Day 13 p = .0082 experimental group survival (P 5.0 x 10e6 = 0.0082). d Mouse model using U937 cells to induce Days after injection measurable tumor formation. Mice treated with either 123b- 33bcCAR T-cells or control T- D. IVIS imaging of U937-Luc injected murine model E. U937 tumor burden cells. Tumor burden visualized on days 3, 6, 9, and 13 (left panel) with e graphical Control T-cells 123-33cCAR T-cells 7.0 x 10e4 representation on the right. f Log-rank mantel-cox test shows Day 3 ** * signiﬁcance for improved 7.0 x 10e3 experimental group survival (P 2.0 x 10e5 = 0.0082) Day 6 Days after injection 2.0 x 10e4 Survival Proportions: U937 mice F. 6.0 x 10e5 Day 9 Control T-cells cCAR T-cells 6.0 x 10e4 6.0 x 10e6 p = .0082 Day 13 6.0 x 10e5 Days after injection cells. These ﬁndings correlated with observed robust anti- absent (<1% tumor burden) in 123b-33bcCAR-treated Jur- leukemia activity and supported overall signiﬁcantly katxp123 mice, and Jurkatxp33 mice had only 10% residual improved survival. tumor. 123b-33bcCAR T-cells display discrete antigen 123b-33bcCAR T-cells can be rapidly depleted by targeting ability in vivo alemtuzumab in vivo + + To assay speciﬁc CD123 or CD33 cell depletion and The development of safety methods to eliminate 123b- verify compound scFv efﬁcacy, two additional xenograft 33bcCAR T-cells from AML patients after tumor depletion mouse models of artiﬁcial tumor cells independently may be necessary in emergencies due to unexpected side expressing each antigen (Jurkatxp123 or Jurkatxp33) were effects of cCAR therapy. T-cells express CD52 on the cell generated (Fig. 5). Both groups were treated with 123b- surface and a CD52-speciﬁc antibody, CAMPATH (alem- 33bcCAR and control T-cells to test 123b-33bcCAR’s tuzumab), can eliminate CD52 cells from circulation. independent antigen targeting ability. There was a sig- To evaluate the feasibility of an alemtuzumab-mediated niﬁcant difference (P < 0.002) in IVIS measurement of depletion strategy as a natural safety-switch for 123b- tumor burden between the control group and the 123b- 33bcCAR T-cells, we established a mouse model with a 33bcCAR treatment group from day 7 onwards for the systemic injection of 123b-33bcCAR T-cells. Mice were Jurkatxp123 mice and as early as day 3 for the Jurkatxp33 treated with alemtuzumab and the tissues were analyzed as mice (Fig. 5b, e). By day 18, tumor cells were virtually described in the workﬂow in Fig. 6a. After 6 h of Percent Survival Residual Tumor Percent Survival Residual Tumor Anti-CD123 and CD33 compound CAR T-cells target acute myeloid 10 leukemia 1323 123b-33bcCAR T-cells suppress Jurkat xp123 and Jurkat xp33 tumor growth in vivo A. IVIS imaging of Jurkat xp123-Luc injected murine model B. Jurkat xp123 tumor burden Dorsal Ventral Control Control cCAR Control cCAR cCAR 1.0 x 10e5 ** * * Day 3 1.0 x 10e4 1.0 x 10e5 Day 7 1.0 x 10e4 1.0 x 10e5 Days after injection Day 10 1.0 x 10e4 Survival Proportions: Jurkat xp123 mice C. 3.0 x 10e5 Control T-cells Day 14 123b-33bcCAR T-cells 3.0 x 10e4 2.0 x 10e6 Day 18 2.0 x 10e5 Days after injection D. IVIS imaging of Jurkat xp33-Luc injected murine model E. Jurkat xp33 tumor burden Dorsal Ventral Control cCAR Control cCAR Control *** * * 1.0 x 10e5 cCAR Day 3 1.0 x 10e4 1.0 x 10e5 Day 7 1.0 x 10e4 1.0 x 10e5 Days after injection Day 10 Survival Proportions: Jurkat xp33 mice 1.0 x 10e4 F. 3.0 x 10e5 Control T-cells Day 14 123b-33bcCAR T-cells 3.0 x 10e4 2.0 x 10e6 Day 18 2.0 x 10e5 Days after injection Fig. 5 123b-33bcCAR T-cells display discrete antigen targeting in two d Mouse model using Jurkatxp33 cells to induce measurable tumor in vivo xenograft mouse models. a Mouse model using Jurkatxp123 formation. Mice treated with either 123b-33bcCAR T-cells or control cells to induce measurable tumor formation. Mice treated with either T-cells. Tumor burden visualized on days 3, 7, 10, 14, and 18 (left 123b-33bcCAR T-cells or control T-cells. Tumor burden visualized on panel) with e graphical representation on the right. f) Kaplan–Meier days 3, 7, 10, 14, and 18 (left panel) with b graphical representation on survival curve represents survival outcomes the right. c Kaplan–Meier survival curve represents survival outcomes. alemtuzumab administration, ﬂow cytometry analysis con- alemtuzumab as a natural safety-switch to deplete 123b- ﬁrmed ~90% depletion of 123b-33bcCAR T-cells in the 33bcCAR T-cells from circulation. circulating peripheral blood (Fig. 6b, Supplementary Fig. S5A), and the depletion rate increased to >98% 48 h post- alemtuzumab (Fig. 6c, Supplementary Fig. S5A). Further- Discussion more, 5 days post-alemtuzumab, the 123b-33bCAR T-cells were virtually eradicated from the peripheral blood, spleen, While initial remission rates of ~90% are commonly seen in liver, and bone marrow (Supplementary Fig. S5B). Alem- CD19CAR-treated B-ALL patients, most relapse within a tuzumab, therefore, rapidly and efﬁciently eliminated both year. The relapse may arise from incomplete chimeric antigen peripheral blood cCAR T-cells as well as tissue-inﬁltrating coverage resulting in antigen escape. A single target for CAR- cCAR T-cells. These ﬁndings support the use of based treatment may, therefore, not be sufﬁcient to prevent Day 3 Day 7 Day 10 Day 14 Day 18 Day 3 Day 7 Day 10 Day 14 Day 18 Percent Tumor Percent Survival Percent Tumor Percent Survival 1324 J. C. Petrov et al. Depletion of infused 123b-33bcCAR T-cells following treatment with alemtuzumab 6 h Peripheral blood 24 h Peripheral blood Negative Control w/o Alemtuzumab with Alemtuzumab Negative Control w/o Alemtuzumab with Alemtuzumab CD45 CD45 Fig. 6 Depletion of infused 123b-33bcCAR T-cells following treat- persistence in peripheral blood 6 h later. Presence of 123b-33bcCAR ment with alemtuzumab. a Experimental schema to evaluate the effect T-cells was detected by ﬂow cytometry. c Representation of persis- of alemtuzumab administration after 123b-33bcCAR T-cells infusion tence of infused 123b-33bcCAR T-cells in peripheral blood 24 h later. into NGS mice. b Representation of 123b-33bcCAR T-cells Presence of 123b-33bcCAR T-cells was detected by ﬂow cytometry leukemia relapse and more effective CAR T-cell treatments to overexpress CD123. While current therapies aim to de-bulk prevent relapse are urgently needed. Single-CAR treatment of the disease by logs of depletion of these cells, they fail to AML poses similar concerns for antigen escape, emphasizing eliminate LSCs, thus allowing the regeneration of disease the importance of dual targeting approaches. and blasts. A more successful strategy is to simultaneously CD123 and CD33 are good cCAR targets as virtually all target the roots of the disease (CD123 + CD34 + CD38- AML cells express one or both of these antigens. We LSCs) and bulk of the disease (CD33 + blasts). developed a cCAR targeting both CD123 and CD33 anti- A variety of approaches have been used to target either gens simultaneously as a strategy for more comprehensive CD123 or CD33 in AML patients, including antibody-drug disease ablation. Dual targeting of AML by a combined conjugates and T-cell recruiting antibody constructs . − − CD123 CD33 directed therapy may overcome the pitfalls These approaches have met limited success with high rates of single-antigen therapy by preventing relapse due to of relapse, which may be due to an inability to simulta- antigen loss. While loss of a single antigen under antigen- neously target bulk disease and LSCs. Several clinical trials speciﬁc selection pressure is possible, loss of two antigens have employed CARs to individually target either CD123 simultaneously is much more unlikely. (NCT02159495, NCT03190278) or CD33 (NCT01864902; One important prognostic indicator of AML is minimal NTC0186902; NCT02944162; NCT03126864); although residual disease, deﬁned as small numbers of leukemic cells sparse preliminary data supports limited success. In a small which remain after treatment. This population is pre- clinical trial using CD33 CAR T-cells to treat one patient dominantly comprised of LSCs and is able to escape con- with refractory AML, CD33 blast cells were decreased at ventional chemotherapy, resulting in relapse . Puriﬁed two weeks, but had returned to high numbers by week 9 CD34 + CD38 leukemic cells are essentially LSCs that after treatment, accompanied by undesirable side effects CD3 CD3 Anti-CD123 and CD33 compound CAR T-cells target acute myeloid 10 leukemia 1325 . The same group also treated a single patient with anti- expanded 123b-33bcCAR T-cells at these endpoints, sug- CD123 CAR T-cells, and reported no clear anti-leukemic gesting that each unit is fully functional and independently effect. Although not enough data from these trials have been sufﬁcient. published to draw conclusions, these early results suggest A major concern in translating our 123b-33bcCAR to the that single-antigen targeting might not be sufﬁcient to stop clinic is the potential for “on-target off-tumor” effects. To AML development and disease progression. rapidly eliminate our powerful cCAR T-cells in the clinic, To offer more potent and comprehensive coverage to we employed an alemtuzumab-mediated depletion strategy. eliminate both LSCs and bulk disease and to minimize Within 6 h of low-dose (0.1 mg/kg) alemtuzumab safety- antigen escape, we propose a cCAR therapy combining switch administration, 90% of cCAR T-cells were elimi- both the CD123 and CD33 targets. The FDA has granted nated from murine blood and tissues, with almost no cells this 123b-33bcCAR orphan drug designation for AML remaining at 24 h. Another group showed that 4–6 weeks of treatment (#17-6031). To study the efﬁcacy of the 123b-33b cCAR therapy was sufﬁcient to eliminate leukemic disease compound CAR (123b-33bcCAR), we identiﬁed three cri- in vivo, and subsequent alemtuzumab administration did teria for its feasibility for the treatment of AML. First, its not hamper CAR therapy . basic cytotoxic functionality to ablate AML cell lines and The 123b-33bcCAR offers a great opportunity in the patient tumor samples. Second, the ability of each discrete clinical setting of AML treatment. Since many AML cCAR unit to independently target its antigen and eliminate patients relapse within a short time of induction and pre- a target expressing only one antigen or both antigens, both transplant, the successful eradication of AML blasts and in vitro and in vivo. Third, our ability to rapidly and LSCs by 123b-33bcCAR T-cells may provide a better completely terminate the cCAR in vivo. window of opportunity to treat these patients with allo- We ﬁrst showed that the 123b-33bcCAR had impressive geneic stem cell transplant. As a bridge to transplant, the cytotoxic activity against leukemic cell lines and patient cCAR would be administered at a speciﬁed dose and the tumors, reaching a nearly 100% killing rate. The cCAR disease status would be monitored closely for 4–6 weeks. targeted and responded to a variety of primary leukemia Preparation of allogeneic transplant would begin once no + + samples and efﬁciently ablated CD123 LSCs and CD33 evidence of disease remains. The patient would receive the AML bulk cells in these samples. We further demonstrated conditioning regimen, followed by a low dose of alemtu- the remarkable killing capacity of the cCAR through dose- zumab, and ﬁnally a stem cell transplant. dependent co-cultures reaching effector:target (E:T) cell There is a dire need for a new AML treatment. Current ratios as low as 0.25:1. Low-dose co-cultures with the 123b- clinical trials targeting various antigens via CAR T-cells 33bcCAR T-cells were comparable to higher E:T ratios, hold great promise, but there is limited published data. suggesting that the cCAR technology is incredibly Preliminary results from these trials suggest that single- powerful. antigen approaches may not be enough. As evidenced by We next validated the ability of compound therapy to the current treatment regimens, a means of eliminating the target multiple antigens by independently testing each target LSCs comprising minimal residual disease while also + − - + (CD123 CD33 or CD123 CD33 ). We showed that the ablating the bulk leukemia, effectively uprooting the AML 123b-33bcCAR was able to ablate both the CD123- tree, is essential for long-term results. Our work supports a expressing Jurkatxp123 cells and the CD33-expressing compound CAR as a novel and powerful means of advan- Jurkatxp33 cells independently both in co-culture assays cing current AML therapies. and in mouse models. Mice treated with 123b-33bcCAR Acknowledgements The authors thank Todd Rueb and Rebecca showed trends of improved survival as compared to control- Connor for valuable technical assistance with FACS. They thank treated mice. We also showed the 123b-33bcCAR promoted Amelia E. Firor for her 123b-33bcCAR schematic design. This work sustained in vivo activity against the MOLM13 and U937 was supported by Macau Science and Technology Development Fund cell lines, as well as superior murine survival in both (FDCT-010/2016/A1-SKL). models. The eventual leukemia relapse in the mice may Author contributions Designed the experiments, interpreted the data, have been due to the limited ability of human CAR T-cells and wrote the manuscript: YM. Designed and performed the experi- to effectively penetrate all murine reservoirs. The presence ments, interpreted data, and wrote the manuscript: JCP and MW . Performed experiments: KGP, KHC, XC. Wrote the manuscript: KGP, of murine reservoirs that harbor leukemia and allow relapse LEY, KHC, XS, HL, LL, HS, NH, XJ, FL. have been documented by other groups [17, 23–25]. Moreover, mice bearing xenografts of human malignancies Compliance with ethical standards do not recapitulate the human microenvironment, which allows for 1000- to 10,000-fold CAR T-cell expansion . Conﬂict of interest Yupo Ma is a co-founder of iCell Gene Ther- Importantly, all cCAR-treated mouse hematological tissues apeutics, LLC. were largely free of tumor with large populations of 1326 J. C. Petrov et al. Open Access This article is licensed under a Creative Commons 12. Testa U, Pelosi E, Frankel A. CD 123 is a membrane biomarker Attribution 4.0 International License, which permits use, sharing, and a therapeutic target in hematologic malignancies. Biomark adaptation, distribution and reproduction in any medium or format, as Res. 2014;2:4. long as you give appropriate credit to the original author(s) and the 13. Munoz L, Nomdedeu JF, Lopez O, Carnicer MJ, Bellido M, source, provide a link to the Creative Commons license, and indicate if Aventin A, et al. Interleukin-3 receptor alpha chain (CD123) is changes were made. The images or other third party material in this widely expressed in hematologic malignancies. Haematologica. article are included in the article’s Creative Commons license, unless 2001;86:1261–9. indicated otherwise in a credit line to the material. If material is not 14. Horton SJ, Huntly BJ. Recent advances in acute myeloid leukemia included in the article’s Creative Commons license and your intended stem cell biology. Haematologica. 2012;97:966–74. use is not permitted by statutory regulation or exceeds the permitted 15. Pardanani A, Lasho T, Chen D, Kimlinger TK, Finke C, Zblewski use, you will need to obtain permission directly from the copyright D, et al. Aberrant expression of CD123 (interleukin-3 receptor- holder. To view a copy of this license, visit http://creativecommons. alpha) on neoplastic mast cells. Leukemia. 2015;29:1605–8. org/licenses/by/4.0/. 16. Liu K, Zhu M, Huang Y, Wei S, Xie J, Xiao Y. CD123 and its potential clinical application in leukemias. Life Sci. 2015;122:59–64. References 17. Chen KH, Wada M, Pinz KG, Liu H, Lin KW, Jares A, et al. Preclinical targeting of aggressive T-cell malignancies using anti- 1. Ehninger A, Kramer M, Rollig C, Thiede C, Bornhauser M, von CD5 chimeric antigen receptor. Leukemia. 2017;31:2151–60. Bonin M, et al. Distribution and levels of cell surface expression 18. Chen KH, Wada M, Pinz KG, Liu H, Shuai X, Chen X, et al. A of CD33 and CD123 in acute myeloid leukemia. Blood Cancer J. compound chimeric antigen receptor strategy for targeting multi- 2014;4:e218. ple myeloma. Leukemia. 2017;32:402–12. 2. Bose P, Vachhani P, Cortes JE. Treatment of relapsed/refractory 19. Chen KH, Wada M, Firor AE, Pinz KG, Jares A, Liu H, et al. acute myeloid leukemia. Curr Treat Options Oncol. 2017;18:17. Novel anti-CD3 chimeric antigen receptor targeting of aggressive 3. Percival ME, Estey E. Emerging treatments in acute myeloid T cell malignancies. Oncotarget. 2016;7:56219–32. leukemia: current standards and unmet challenges. Clin Adv 20. Pinz K, Liu H, Golightly M, Jares A, Lan F, Zieve GW, et al. Hematol & Oncol: Hamp;O. 2017;15:632–42. Preclinical targeting of human T-cell malignancies using CD4- 4. Fan M, Li M, Gao L, Geng S, Wang J, Wang Y, et al. Chimeric speciﬁc chimeric antigen receptor (CAR)-engineered T cells. antigen receptors for adoptive T cell therapy in acute myeloid Leukemia. 2016;30:701–7. leukemia. J Hematol & Oncol. 2017;10:151. 21. Lichtenegger FS, Krupka C, Haubner S, Kohnke T, Subklewe M. 5. Firor AE, Jares A, Ma Y. From humble beginnings to success in Recent developments in immunotherapy of acute myeloid leuke- the clinic: Chimeric antigen receptor-modiﬁed T-cells and impli- mia. J Hematol & Oncol. 2017;10:142. cations for immunotherapy. Exp Biol Med (Maywood, NJ). 22. Wang QS, Wang Y, Lv HY, Han QW, Fan H, Guo B, et al. 2015;240:1087–98. Treatment of CD33-directed chimeric antigen receptor-modiﬁed 6. Mikkilineni L, Kochenderfer JN. Chimeric antigen receptor T-cell T cells in one patient with relapsed and refractory acute myeloid therapies for multiple myeloma. Blood. 2017;130:2594–602. leukemia. Mol Ther. 2015;23:184–91. 7. Pollyea DA, Gutman JA, Gore L, Smith CA, Jordan CT. Targeting 23. Mamonkin M, Rouce RH, Tashiro H, Brenner MK. A T-cell- acute myeloid leukemia stem cells: a review and principles for the directed chimeric antigen receptor for the selective treatment of T- development of clinical trials. Haematologica. 2014;99:1277–84. cell malignancies. Blood. 2015;126:983–92. 8. O’Hear C, Heiber JF, Schubert I, Fey G, Geiger TL. Anti-CD33 24. Brentjens RJ, Santos E, Nikhamin Y, Yeh R, Matsushita M, La chimeric antigen receptor targeting of acute myeloid leukemia. Perle K, et al. Genetically targeted T cells eradicate systemic acute Haematologica. 2015;100:336–44. lymphoblastic leukemia xenografts. Clin Cancer Res. 2007;13(18 9. Jilani I, Estey E, Huh Y, Joe Y, Manshouri T, Yared M, et al. Pt 1):5426–35. Differences in CD33 intensity between various myeloid neo- 25. Barrett DM, Liu X, Jiang S, June CH, Grupp SA, Zhao Y. plasms. Am J Clin Pathol. 2002;118:560–6. Regimen-speciﬁc effects of RNA-modiﬁed chimeric antigen 10. Jiang Y, Xu P, Yao D, Chen X, Dai H. CD33, CD96 and Death receptor T cells in mice with advanced leukemia. Hum Gene Ther. Associated Protein Kinase (DAPK) expression are associated with 2013;24:717–27. the survival rate and/or response to the chemotherapy in the 26. Ramos CA, Savoldo B, Dotti G. CD19-CAR trials. Cancer J. patients with Acute Myeloid Leukemia (AML). Med Sci Monit. 2014;20:112–8. 2017;23:1725–32. 27. Tasian SK, Kenderian SS, Shen F, Ruella M, Shestova O, 11. Chevallier P, Robillard N, Ayari S, Guillaume T, Delaunay J, Kozlowski M, et al. Optimized depletion of chimeric antigen Mechinaud F, et al. Persistence of CD33 expression at relapse in receptor T cells in murine xenograft models of human acute CD33(+) acute myeloid leukaemia patients after receiving Gemtu- myeloid leukemia. Blood. 2017;129:2395–407. zumab in the course of the disease. Br J Haematol. 2008;143:744–6.
– Springer Journals
Published: Feb 25, 2018