Research ArticleCancer Immunotherapy

CD19-Targeted T Cells Rapidly Induce Molecular Remissions in Adults with Chemotherapy-Refractory Acute Lymphoblastic Leukemia

See allHide authors and affiliations

Science Translational Medicine  20 Mar 2013:
Vol. 5, Issue 177, pp. 177ra38
DOI: 10.1126/scitranslmed.3005930


Adults with relapsed B cell acute lymphoblastic leukemia (B-ALL) have a dismal prognosis. Only those patients able to achieve a second remission with no minimal residual disease (MRD) have a hope for long-term survival in the context of a subsequent allogeneic hematopoietic stem cell transplantation (allo-HSCT). We have treated five relapsed B-ALL subjects with autologous T cells expressing a CD19-specific CD28/CD3ζ second-generation dual-signaling chimeric antigen receptor (CAR) termed 19-28z. All patients with persistent morphological disease or MRD+ disease upon T cell infusion demonstrated rapid tumor eradication and achieved MRD complete remissions as assessed by deep sequencing polymerase chain reaction. Therapy was well tolerated, although significant cytokine elevations, specifically observed in those patients with morphologic evidence of disease at the time of treatment, required lymphotoxic steroid therapy to ameliorate cytokine-mediated toxicities. Indeed, cytokine elevations directly correlated to tumor burden at the time of CAR-modified T cell infusions. Tumor cells from one patient with relapsed disease after CAR-modified T cell therapy, who was ineligible for additional allo-HSCT or T cell therapy, exhibited persistent expression of CD19 and sensitivity to autologous 19-28z T cell–mediated cytotoxicity, which suggests potential clinical benefit of additional CAR-modified T cell infusions. These results demonstrate the marked antitumor efficacy of 19-28z CAR-modified T cells in patients with relapsed/refractory B-ALL and the reliability of this therapy to induce profound molecular remissions, forming a highly effective bridge to potentially curative therapy with subsequent allo-HSCT.


Despite available chemotherapy and allogeneic hematopoietic stem cell transplantation (allo-HSCT), adult patients with relapsed B cell acute lymphoblastic leukemia (B-ALL) have a very poor prognosis. Long-term survival of adult patients with relapsed B-ALL is dependent upon achieving a complete remission (CR) induced through salvage chemotherapy followed by allo-HSCT (1, 2). Unfortunately, many patients never receive a potential life-saving allo-HSCT due to a failure in achieving a second CR after salvage chemotherapy (1). Further, in patients undergoing an allo-HSCT, those with minimal residual disease (MRD+) by fluorescence-activated cell sorting (FACS) or polymerase chain reaction (PCR) have a significantly worse prognosis compared to patients with no evidence of MRD (MRD) at the time of allo-HSCT (3). For this reason, new therapeutic regimens for this patient population are needed.

A patient’s own T cells may be genetically modified to express an artificial T cell receptor—termed a chimeric antigen receptor (CAR)—designed to be specific to a tumor-associated antigen. We and others have previously reported on promising preclinical outcomes of CAR-modified T cells targeted to the B cell CD19 antigen (47). CD19 is expressed on normal B cells as well as on most B cell malignancies including low-grade chronic lymphocytic leukemias (CLLs), B cell non-Hodgkin’s lymphomas (NHLs), and more aggressive B-ALL. Despite differences in CAR and clinical trial designs, expansion of this technology to treat patients with low-grade B cell malignancies (CLL and follicular lymphoma) at three different centers has demonstrated significant antitumor responses after infusion of CD19-targeted autologous T cells (812). Although promising clinical outcomes have been reported in patients with low-grade B cell tumors, to date, there are no reported clinical outcomes using this CD19-targeted adoptive T cell therapy approach in patients with relapsed B-ALL, a far more aggressive disease with a markedly worse prognosis.

We have treated five relapsed B-ALL adult patients with autologous second-generation CD19-targeted CAR (19-28z) T cells after salvage chemotherapy. We report the marked ability of autologous 19-28z CAR-modified T cells to induce MRD CRs in patients with relapsed and/or chemotherapy-refractory B-ALL. Further, we demonstrate that after T cell infusion, cytokine-mediated toxicities, similar to reported toxicities (912) in low-grade B cell malignancies with CD19-targeted CAR-modified T cells, correlate to the degree of tumor burden at the time of CAR-modified T cell infusion. Our data demonstrate the life-saving potential of this technology for the treatment of relapsed B-ALL.


Infusion of 19-28z CAR-modified T cells induced MRD remissions

Patients with relapsed B-ALL, not previously treated with allo-HSCT, independent of remission status after salvage chemotherapy, were eligible for therapy with autologous 19-28z+ T cells on this clinical protocol (figs. S1 and S2). Patients were treated at a dose of 1.5 × 106 to 3 × 106 autologous 19-28z+ T cells/kg (13) (table S1) after previous conditioning therapy with cyclophosphamide (1.5 to 3.0 g/m2). For ethical reasons, and as per the protocol, after adoptive T cell therapy, eligible patients underwent subsequent allo-HSCT, limiting the time for follow-up observation of CAR-modified T cells and long-term molecular remissions in four of five patients treated to date.

Of the five treated subjects, patients MSK-ALL04 and MSK-ALL05 exhibited persistent chemotherapy-refractory disease with 63 and 70% blast cells in the bone marrow (BM), respectively, after salvage chemotherapy; two other patients had achieved morphologic CRs with evidence of MRD by deep sequencing PCR and FACS (MSK-ALL01 and MSK-ALL06) (Table 1). Subsequent to adoptive 19-28z+ T cell therapy, all patients were MRD as assessed by deep sequencing PCR, demonstrating the loss of detectable malignant clone immunoglobulin H (IgH) rearrangements. Of those patients with morphologic evidence of disease at the time of modified T cell therapy, MSK-ALL04 achieved a morphologic CR by day 11 and MRD status by day 59; MSK-ALL05 markedly achieved both a morphologic CR and MRD status within 8 days of T cell infusion (Fig. 1 and Table 2). MRD+ patient MSK-ALL01 achieved MRD status by day 28, whereas MRD+ patient MSK-ALL06 was MRD by deep sequencing on day 30 and remained with MRD disease up to the time of allo-HSCT 122 days after therapy (Table 2). Collectively, all patients achieved an MRD status as the optimal disease response after 19-28z+ T cell infusion. Assessment of the durability of MRD responses in this study was limited by patients subsequently undergoing allo-HSCT 1 to 4 months after therapy in four of five patients with the longest MRD status reported in this cohort out to 122 days after CAR-modified T cell therapy.

Table 1

Patient characteristics and response summary. Cy, cyclophosphamide; Vinc, vincristine; Pred, prednisone; Etop, etoposide; Peg, pegylated asparaginase; Mito, mitoxantrone; CR1, first complete remission; MRD, minimal residual disease as assessed by deep sequencing (see Supplementary Materials); Allo-SCT, allogeneic stem cell transplant; FISH, fluorescence in situ hybridization. All patients were treated with cyclophosphamide before T cell infusion, either 1.5 g/m2 (MSK-ALL04, MSK-ALL05, and MSK-ALL06) or 3.0 g/m2 (MSK-ALL01 and MSK-ALL03).

View this table:
Fig. 1

Rapid antitumor effects mediated by 19-28z T cells. (A) BM aspirates before and after treatment with 19-28z T cells in two patients with morphologic chemotherapy-refractory B-ALL. Cyclophosphamide was given at day −1, and CD19 CAR-targeted T cells were infused on days 1 and 2. Left panels: BM before CAR-modified T cell therapy demonstrated predominant blast cells with an absence of normal BM precursors. For MSK-ALL04, the left panel includes an inset with ×100 magnification. Middle panels: BM aspirates done shortly after 19-28z T cell infusion and hypocellular with normal stromal elements, histiocytes, and no evidence of blasts. Right panels: By 1 to 2 months after CAR-modified T cell therapy, there is BM recovery with normal hematopoiesis and no evidence of abnormal blasts. (B) Flow cytometry for CD19 and CD10 expression in BM before and after treatment. Cells were gated on CD45+7AAD cells.

Table 2

Deep sequencing for IgH rearrangements before and after CD19 CAR-targeted T cell therapy. Adaptive Biotechnologies performed multiplex PCR and deep sequencing on genomic DNA prepared from BM aspirated on the noted day (see Supplementary Materials for further detail). Malignant IgH rearrangement refers to IgH rearrangements associated with the B-ALL tumor cells. Total numbers of IgH rearrangements are derived from both malignant and nonmalignant B cells.

View this table:

One patient, MSK-ALL04, ineligible for either allo-HSCT or additional CAR T cell therapy, relapsed 90 days after treatment in the setting of previous high-dose steroid therapy for cytokine-mediated toxicities, which may in turn have abrogated CAR-modified T cell persistence, enhancing the ability of the B-ALL CD19+ tumor cells to escape immune surveillance mediated by the infused CAR-modified T cells. Despite this isolated case of relapsed disease, the presented clinical outcomes on this cohort of patients collectively, for the time, demonstrate the profound clinical benefit of CD19-targeted CAR technology in the setting of relapsed adult B-ALL, a highly aggressive and predominantly fatal condition (1, 2).

Cytokine-mediated toxicities correlated to tumor bulk at the time of CAR T cell infusion

Therapy with CD19-targeted CAR-modified T cells has reported toxicities of high fevers, hypotension, and elevated proinflammatory serum cytokines in patients with CLL and B cell NHL (9, 10, 12). For this reason, serum cytokine levels were serially monitored in all treated patients to correlate cytokine levels to subsequent treatment-related toxicities and patient clinical outcomes (table S2). Patients with evidence of persistent disease at the time of CAR-modified T cell infusions generally showed early proinflammatory cytokine responses starting 3 to 5 days after modified T cell infusion, albeit at markedly different levels. The highest cytokine elevations, including soluble interleukin-2 receptor α (sIL-2Rα), interferon-γ (IFN-γ), IL-6, and IFN-inducible protein 10 (IP10), were seen in MSK-ALL04 and MSK-ALL05, the two patients with the highest tumor burden at the time of T cell infusion (Fig. 2A and fig. S3); these two patients also had the largest number of detectable CD19 CAR-targeted T cells (Fig. 2B and table S3). Cytokine elevations were far more modest or undetectable in the other MRD+ and MRD patients (Fig. 2A). The degree of cytokine elevation thus correlated to the bulk of residual disease at the time of adoptive T cell infusion (Fig. 2C) and was coincident with post-infusion fevers and episodes of relative hypotension (Fig. 3). Three of five patients, all with the highest levels of measurable disease, experienced transient fevers after adoptive T cell therapy (Fig. 3 and table S2). Increased fever severity and persistence accompanied by relative hypotension and transient mental status changes were observed in MSK-ALL04 and MSK-ALL05 (Fig. 3 and table S2). Both MSK-ALL04 and MSK-ALL05 received high-dose lymphotoxic steroid therapy starting day 6 followed by a slow taper, which rapidly ameliorated constitutional symptoms (Fig. 3) and concomitantly normalized serum cytokine levels (Fig. 2A). Our data present a positive correlation between cytokine surges with associated clinical toxicities in patients treated with CD19-targeted CAR T cells to the degree of tumor burden at the time of CD19-targeted T cell infusions (Fig. 2C).

Fig. 2

Cytokine release and T cell persistence were increased in patients with high tumor burdens. (A) Pre- and posttreatment serum from patients was evaluated for listed cytokines (pg/ml). The asterisk marks initiation of steroids for MSK-ALL04 and MSK-ALL05. MSK-ALL04 started dexamethasone (20 mg) every 8 hours on day 6, then held for 1 day, and restarted on day 8. Dexamethasone was then tapered off over 2 weeks. MSK-ALL05 was started on dexamethasone (20 mg) every 12 hours, and it was tapered off over 12 days. (B) 19-28z T cells in the blood were detected by quantitative PCR as described (8). From these results, absolute 19-28z T cell counts were calculated. MSK-ALL04 is missing a time point immediately before administration of steroids. (C) Correlation between the maximum IFN-γ (pg/ml) and tumor burden for each patient (left panel) and correlation between the maximum 19-28z T cell count and tumor burden for each patient were calculated as the Spearman rank correlation coefficient (r), which is listed on both panels. Tumor burden is the number of malignant IgH clonotypes identified in the pretreatment BM. Tumor burden, cytokines, and 19-28z T cell counts were rank-ordered to calculate the correlation coefficient. The Spearman rank correlation coefficients for tumor burden and maximum IP10 (r = 0.91), IL-2 (r = 0.88), and IL-6 (r = 0.72) were also calculated.

Fig. 3

Persistent fevers in patients with high tumor burden after infusion with 19-28z CAR+ T cells. The maximum temperature (°C) in a 24-hour period is noted for all patients. Days listed range from day −1 (cyclophosphamide) to 14 days after CD19 CAR-targeted T cell infusion. The asterisk marks day 6 when MSK-ALL04 and MSK-ALL05 were both started on high-dose steroids. The green line marks the minimum temperature for a fever (38°C).

CD19-targeted 19-28z+ T cells exhibited substantial in vivo expansion and persistence

In vivo expansion and persistence of 19-28z+ T cells over time are likely to be critical variables with respect to clinical outcomes. To this end, we assessed both expansion and persistence of CAR-modified T cells by FACS and PCR in this cohort of patients. Overall, CAR-modified T cells were detectable by FACS or reverse transcription PCR (RT-PCR) in the blood and BM 3 to 8 weeks after 19-28z+ T cell infusion (Figs. 2B and 4). In vivo expansion, as measured by peak levels of detectable CAR-modified T cells, additionally positively correlated to tumor burden at the time of modified T cell infusion (Fig. 2C). Analysis of T cell persistence in this study is compromised in part by the fact that eligible patients undergo allo-HSCT relatively soon after infusion with CAR-modified T cells (1 to 4 months). Analyses of long-term CAR-modified T cell persistence in MSK-ALL04 and MSK-ALL05 were additionally compromised by prolonged high-dose steroid therapies to treat cytokine-mediated toxicities, which coincided with a rapid drop in CAR-modified T cell counts and clinical evidence of cytokine-meditated toxicities (Fig. 2B). However, although loss of malignant clones by deep sequencing PCR was uniformly noted in all patients with persistent disease at the time of 19-28z+ T cell therapy, we concomitantly observed recovery of normal B cell clones in all patients, consistent with waning persistence of CAR-modified T cells and, importantly, recovery of normal B cell lymphopoiesis (Table 2). Notably, MSK-ALL05, who rapidly became MRD (by day 8), exhibited a remarkable rapid neutrophil recovery in the context of chronic daily granulocyte colony-stimulating factor infusions, despite prolonged neutropenia before T cell therapy, demonstrating the preservation and rapid restoration of normal hematopoiesis in the context of CD19-targeted 19-28z+ T cell therapy (fig. S4).

Fig. 4

19-28z T cells could be detected in the blood of treated patients. Within 1 week of 19-28z T cell infusion, peripheral T cells were isolated from the blood and activated ex vivo as described in the Supplementary Materials and as reported elsewhere (8). Reactivated T cells were evaluated by flow cytometry for CD3 and 19-28z CAR expression. Cells displayed within the FACS plots have been gated as CD45+7AAD.

Durable remissions occurred after treatment with CD19-targeted 19-28z+ T cells and allo-SCT

Of the five patients treated on this protocol, four have undergone subsequent allo-HSCT (Table 1). MSK-ALL01 died of a suspected pulmonary embolus 2 months after allo-SCT while in CR without any evidence of disease. The other three patients have had no significant complications after allo-SCT, and all remain in CR. MSK-ALL03 remains in CR 18 months after allo-SCT. MSK-ALL05 is in CR 3 months after allo-SCT. MSK-ALL06 was in MRD remission for 122 days after T cell infusion before being treated with an allo-SCT. To date, this patient remains MRD 6 weeks after allo-SCT.

Relapsed B-ALL disease after 19-28z+ T cell therapy retained CD19 expression as well as sensitivity to CAR T cell–mediated lysis

MSK-ALL04, who was ineligible for allo-HSCT because of multiple, preexisting comorbidities, was observed expectantly after CAR-modified T cell therapy. This patient relapsed 13 weeks after T cell therapy and subsequently expired (Table 1). The relapsed tumor cells expressed the same malignant IgH rearrangement identified in the initial malignant clone and expressed the target CD19 antigen at pretreatment levels (Fig. 5A). Further, the recurrent tumor cells retained sensitivity to lysis by autologous 19-28z+ T cells as assessed by a 51Cr release assay (Fig. 5B). The relapse observed in this patient, despite previously having achieved MRD status, is therefore not due to antigen escape but may be imputed, at least in part, to the abrogated persistence of the infused 19-28z+ T cells caused by the requisite need for high-dose steroid therapy to treat cytokine-mediated toxicities (Fig. 2B).

Fig. 5

Relapsed B-ALL tumor cells from MSK-ALL04 retained CD19 expression and sensitivity to 19-28z T cell–mediated killing. (A) CD19 and CD10 expression of B-ALL tumor cells from the initial diagnostic whole-blood sample (left panel) and the post–19-28z relapsed sample. Displayed cells were gated on CD45+7AAD cells. (B) Untransduced (UNT) T cells from the leukapheresis product or 19-28z T cells from the end-of-product formulated cells were incubated with the post–19-28z relapsed B-ALL tumor cells. Effectors were incubated with tumor cells (radiolabeled with 51Cr) at a 36:1 effector/target ratio for 4 hours. 51Cr release was measured and calculated as a killing efficiency as described (8).


Adult patients with relapsed B-ALL have an overall poor prognosis. Those patients who relapse after multiagent chemotherapy protocols have an even more dismal prognosis. The standard of care for these patients is salvage chemotherapy followed by allo-HSCT if in remission and clinically eligible to undergo allo-HSCT. However, many such patients fail to achieve a second CR and/or are ineligible for additional transplantation therapy (1).

Treatment of patients with relapsed indolent B cell malignancies, including CLL and B cell follicular lymphomas, using autologous T cells targeted to the CD19 antigen through the retroviral or lentiviral introduction of a CD19-specific CAR has demonstrated very promising clinical responses (812). Despite a growing body of clinical evidence for efficacy using CAR T cell technology in these indolent B cell tumors, it remains unknown whether this same technology may similarly exhibit significant antitumor efficacy in adults with far more aggressive and often fatal relapsed B-ALL.

The limitations of our study include the short follow-up after T cell infusion as well as the relatively small number of patients treated to date with the 19-28z T cells. The treatment of these patients with allo-SCT, the standard of care for relapsed B-ALL patients, after 19-28z T cell infusion, makes it indeed difficult to discern the curative or long-term remission potential of CD19 CAR-targeted T cells. However, it is unequivocal that patients who were ineligible for allo-SCT and had no further therapeutic option in the face of rapidly progressing refractory disease were promptly brought into profound molecular remission by a single infusion of 19-28z T cells. Therefore, despite the limitations, our work strongly supports the efficacy of 19-28z T cells in adults with acute leukemia.

Here, we summarize the outcomes of five adults with relapsed B-ALL, four of which demonstrated persistent disease after salvage chemotherapy at the time of CAR T cell infusion. Overall, we report that in all patients, ranging from overt morphologic disease to MRD, treatment with autologous T cells uniformly resulted in MRD CRs independent of tumor burden at the time of T cell therapy. These results demonstrate a potent antitumor efficacy of autologous T cells genetically modified to target the CD19 antigen expressed on adult B-ALL tumor cells. These consistent clinical outcomes are meaningful for several reasons: First, these data demonstrate the potential and rapid kinetics of 19-28z CD19-targeted T cells in eradicating chemotherapy-resistant adult B-ALL, and second, the ability to rapidly and reliably achieve MRD status markedly improves the prognosis of these patients by providing a bridge to allo-HSCT under optimal conditions, the standard of care for these patients. Specifically, in the absence of CAR-modified T cell therapy, two patients treated on this cohort would not have been eligible for allo-HSCT, whereas an additional two patients would have undergone this therapy in the MRD+ setting with predicted poor clinical outcomes (3).

Additionally, our studies demonstrate that cytokine-mediated toxicities, which are a significant previously published impediment to the application of this therapy, correlate to tumor bulk at the time of modified T cell infusion. Specifically, we demonstrate an improved side effect profile with CAR-modified T cells if infused in the context of a lower disease burden. These findings suggest that optimal safety of modified T cell infusions is either in the MRD+ setting or perhaps soon after initial salvage chemotherapy in the context of a largely aplastic BM. We find that a T cell–mediated cytokine release syndrome is not requisite to achieving optimal CAR T cell–mediated antitumor efficacy, resulting in MRD CRs of disease.

Sufficient persistence of CD19-targeted CAR-modified cells to achieve tumor eradication (MRD status) is, in theory, essential to optimal cancer antitumor efficacy. However, optimal anti–CD19-targeted T cell therapy as presented by collaborators at the University of Pennsylvania (UPenn) and other institutions presents a balanced analysis with respect to the risks of potentially life-threatening cytokine storms and lifelong B cell aplasias (9, 10, 12). Here, we observed a rapid onset of cytokine elevations in parallel with in vivo T cell expansion correlating to T cell clearance 3 to 8 weeks after infusion as assessed by FACS, RT-PCR, and in vitro expansion assays. Steroid therapy to ameliorate cytokine-driven toxicity in two patients markedly reduced modified T cell populations at a time of maximum detectable CAR-modified T cells (Fig. 2C) similar to a previously reported CLL patient at the UPenn treated with high-dose steroid therapy to control T cell–mediated cytokine toxicities after infusion with CD19-targeted T cells (9). Nonetheless, in our studies, despite the steroid-limited duration of 19-28z CAR-modified T cells, CD19-targeted modified T cell persistence was sufficient to convert all patients with detectable disease to an MRD status (Table 2). Further studies of modified T cell persistence in patients moving on to allo-HSCT were precluded by this additional standard of care therapy. Patient MSK-ALL04, of note, was managed expectantly because of comorbidities prohibiting allo-HSCT and infusion of additional 19-28z CAR-modified T cells. The patient, followed by modified CAR T cell infusion, required additional lymphotoxic steroid therapy soon after modified T cell infusion (day 6), curtailing 19-28z+ T cell persistence and ultimately resulting in the relapse of the patient disease. Notably, in vitro studies demonstrate that the relapsed clone retained CD19 expression and sensitivity to 19-28z+ T cells ex vivo. These results are consistent with the potential of lengthening molecular remissions, converting relapsed patients back to an MRD status, or ultimately fully eradicating CD19+ malignant tumor clones in those patients otherwise ineligible for allo-HSCT through additional infusions of CAR-modified T cells. Considering the decreased tumor burden at the time of second-line therapy with CAR-modified T cell infusion, the latter is likely to induce more modest cytokine elevations and therefore to be better tolerated.

The required dose of CAR-modified T cells was successfully generated in all patients within 2 weeks. To date, we have enrolled a total of 14 patients to this phase 1 protocol. Of these 14 patients, 2 are in CR and are not eligible for T cell infusion at this time, 3 deferred T cell infusion to go directly to allo-SCT, 3 died during salvage chemotherapy, and 5 patients were successfully infused. One patient was lost to follow-up after enrollment. On an intent-to-treat basis, we have generated eight T cell products for enrolled patients including the five described in this study and one who went straight to allo-SCT and two who died during salvage chemotherapy.

The gene transfer efficiency for 19-28z ranged from 5 to 14.2% (table S1). Although there is no clear etiology for this modest gene transfer efficiency, it may be in part due to the nature of acute leukemia or the numerous cycles of high-dose lymphotoxic chemotherapeutic agents received by the patients before T cell harvest. Regardless, it does not appear that the lower CAR transduction efficiency inhibited the antileukemia effect of the 19-28z T cells.

Although we and others have previously reported clinical trial outcomes of CAR T cells targeted to either CD19 or CD20 in patients with low-grade lymphomas and CLL (811, 14, 15) [summarized in (16)], here, we present highly effective clinical outcomes with CD19-targeted CAR T cells in adults in the setting of far more aggressive B-ALL.

Recently published reports from the UPenn demonstrate promising clinical outcomes in CLL patients treated with a second-generation CAR derived from initial studies published elsewhere (17) containing the 4-1BB costimulatory signaling domain. In these studies, investigators demonstrate marked antitumor efficacy, profound cytokine-mediated toxicity, and long-term persistence of both B cell aplasias and autologous T cells expressing this 4-1BB–containing CAR (9, 12). Although results in adult ALL have not yet been reported with 4-1BB/CD3ζ CARs, some differences between the kinetics and the magnitude of the cytokine response between the two CARs are noteworthy. In CLL patients, the 4-1BB/CD3ζ CAR yielded delayed peak cytokine responses seen greater than 10 days out from infusion, in contrast to a more immediate cytokine release (days 3 to 5) seen in our CD28/CD3ζ-containing 19-28z CARs. Additionally, chronic B cell aplasias were associated with long-term persistence of the 4-1BB/CD3ζ T cells, in contrast to the recovery of normal B lymphopoiesis we report here (Table 2). The more limited expansion and persistence of 19-28z+ T cells enable normal B cell recovery over time in the context of a more modest cytokine-mediated toxicity profile. These findings are consistent with murine studies reported so far, which suggest a slower kinetic but ultimately higher accumulation and persistence of 4-1BB T cells that occur without antigen. In xenogeneic NSG chimeras, Milone et al. (7) reported similar therapeutic efficacy of CD28/CD3z and 4-1BB/CD3z T cells in treating CD19+ leukemia, with the latter accumulating to higher levels in delayed fashion. In our own preclinical studies (18), 4-1BB/CD3z T cells were not as effective as CD28z T cells in treating B-ALL in scid/beige xenochimeras (19). Collectively, our studies treating patients with the CD28-containing 19-28z+ CAR-modified T cells provide an alternate approach to using CAR technology, one in which multiple (perhaps two or three) infusions of modified T cells induce CRs in the context of lower tumor burdens, which in turn carry a diminished risk for cytokine-mediated toxicities and long-term B cell aplasias.

Adults with relapsed chemotherapy-refractory B-ALL have a dismal prognosis outside the context of an allo-HSCT. We report the successful induction of MRD remissions in adults with chemotherapy-refractory B-ALL by treatment with CD19-targeted CAR-modified T cells. We acknowledge limited posttreatment follow-up due to the fact that four of five patients subsequently received standard of care therapy with allo-HSCT as stipulated in the protocol and were therefore removed from further follow-up on this protocol. The fifth patient treated was ineligible for allo-HSCT and relapsed 3 months after therapy, although this relapse may in part have been precipitated by high-dose lymphotoxic steroid therapy to treat T cell–mediated cytokine storms soon after modified T cell infusion, which in turn limited persistence of 19-28z CAR-modified T cells. Further, these results are consistent with the potential of lengthening molecular remissions or alternatively reconverting relapsed patients to MRD status through additional infusions of 19-28z CAR-modified T cells. Whether additional infusions with CAR-modified T cells in this particular patient may have changed the clinical outcome remains to be and will be investigated in future relapsed refractory B-ALL patients similarly ineligible for additional therapy with allo-HSCT. The rapid kinetics of 19-28z T cells may prove to be an essential asset in patients with lower tumor burdens but at high risk for an overt relapse. This approach provides a bridge for patients otherwise either ineligible or eligible under very suboptimal conditions (MRD+) to receive potentially life-saving therapy with an allo-HSCT. Furthermore, this therapy has the potential to decrease the relapse rate after allo-HSCT, which can occur in up to one-third of patients (20, 21). In addition, repeated infusions of 19-28z+ T cells in patients unable to undergo allo-HSCT may prolong remission. On the basis of the presented data, treatment of relapsed chemotherapy-refractory B-ALL in adults with 19-28z+ autologous T cells is a very promising approach worthy of continued study. Finally, considering these clinical outcomes, the addition of 19-28z CAR-modified T cells for inclusion into currently up-front adult B-ALL treatment protocols warrants serious consideration.

Materials and Methods

Clinical protocol design

This is a clinical trial designed to assess the safety of infusing autologous T cells modified to express the CD19-specific CAR 19-28z into patients with relapsed B-ALL ( NCT01044069). Patients presenting in CR1 (first CR) or with relapsed disease are eligible for enrollment (figs. S1 and S2). However, patients enrolled during CR1 are only treated if they develop relapsed disease (fig. S2). Patients with relapsed disease receive a chemotherapy regimen chosen by their treating physician and are treated upon either count recovery or evidence of persistent disease. Patients receive high-dose cyclophosphamide chemotherapy (1.5 to 3.0 g/m2) (day −1) followed by a split-dose infusion of CAR-modified T cells on days 1 and 2. The primary endpoint of this study is safety of CAR-modified T cell infusion. Secondary endpoints include studies to detect 19-28z+ T cell persistence, as well as assays to assess morphologic, cytogenetic, and molecular antitumor responses.

Patient clinical background

Five patients with relapsed B-ALL were treated to complete the first cohort of patients on this clinical trial infused with the lowest planned dose of 19-28z CAR-modified T cells (1.5 × 106 to 3.0 × 106 CAR+ T cells/kg) (Table 1 and table S1).

MSK-ALL01, a 66-year-old male with normal cytogenetics, received induction chemotherapy with high-dose cytarabine and mitoxantrone achieving CR1. After three cycles of consolidation chemotherapy, the patient was enrolled on trial with subsequent collection of T cells by leukapheresis. After peripheral blood mononuclear cell (PBMC) collection, the patient was found to have relapsed disease with 50% blasts in the BM. The patient subsequently achieved a second morphological remission with MRD+ after reinduction chemotherapy with prednisone, vincristine, and pegylated asparaginase. Upon BM recovery, the patient received high-dose cyclophosphamide (3 g/m2) followed by split-dose infusion of 19-28z+ T cells. He was referred for an allogeneic SCT while MRD and was removed from study.

MSK-ALL03, a 56-year-old male diagnosed with B-ALL with normal cytogenetics, achieved a CR1 that persisted throughout hyper-CVAD chemotherapy (22) but relapsed during maintenance therapy with 6-mercaptopurine and methotrexate. The patient was subsequently treated on protocol with inotuzumab ozogamicin (23), achieving a very brief remission of disease. Upon relapse, the patient was enrolled on the CAR-modified T cell trial, leukapheresed in the setting of relapsed disease, and achieved a third CR after salvage therapy with prednisone, vincristine, and pegylated asparaginase. Upon BM recovery, the patient was treated with high-dose cyclophosphamide (3 g/m2) followed by split-dose 19-28z+ T cells. He was referred for an allo-SCT while MRD and removed from study.

MSK-ALL04, a 59-year-old male diagnosed with high-risk B-ALL [white blood cell (WBC) count of 147,800/μl at diagnosis] with a 9p21 deletion (9p21−), received initial induction chemotherapy on protocol ECOG2993 (24), achieving a CR after induction phase 1 but presented with relapsed disease after induction phase 2. The patient was enrolled on study, leukapheresed in relapse (WBCs, 54,000/μl), and was subsequently reinduced with a combination of vincristine and prednisone. Upon completion, the patient exhibited persistent disease with 63% blasts present in a hypocellular BM. The patient subsequently received high-dose cyclophosphamide (1.5 g/m2) (day −1) followed by split-dose infusion of 19-28z CAR-modified T cells on days 1 and 2. Long-standing comorbidities (poorly controlled diabetes, renal insufficiency, and coronary artery disease) precluded an allo-SCT, so he was monitored expectantly.

MSK-ALL05 is a 58-year-old male diagnosed with B-ALL and was induced into CR1 on the ECOG2993 regimen (24). He was treated with both phases of induction and an intensification phase but relapsed with 48% blasts in the BM. Fluorescence in situ hybridization (FISH) and cytogenetic analysis of his relapsed B-ALL tumor cells demonstrated that they were hyperdiploid, including an extra copy of the MLL gene, and had homozygotic deletion of the p16 gene. He was enrolled, leukapheresed, and treated with high-dose cytarabine (3 g/m2) and mitoxantrone (40 mg/m2), but his disease persisted with 70% blasts in a hypocellular BM. He was treated with cyclophosphamide (1.5 g/m2) at day −1 followed by split-dose infusion of 19-28z T cells on days 1 and 2. He was referred for an allo-SCT while MRD and subsequently removed from study after transplant.

MSK-ALL06 is a 23-year-old female with relapsed B-ALL (normal cytogenetics). After diagnosis, she was induced into CR1 with a NYII induction regimen (25). She completed all her consolidation and maintenance treatments but relapsed about 9 months after her last maintenance treatment. She was treated with a modified NYII consolidation I regimen (25) with the goal of inducing a CR2 and referring her to an allo-SCT. However, posttreatment analyses confirmed MRD (0.14% blasts in BM), and she developed a small bowel obstruction and underwent a bowel resection, which demonstrated Mucor colitis. She was then enrolled in our trial, leukapheresed, and infused with 19-28z T cells after cyclophosphamide (1.5 g/m2). After her Mucor infection resolved, she underwent an allo-HSCT in an MRD CR at day 122 after CAR-modified T cell therapy.

Generation of 19-28z CAR-modified T cells

PBMCs were obtained from enrolled patients by leukapheresis, and CAR-modified T cells were produced as described (8, 13). Briefly, leukapheresis product was washed and cryopreserved. T cells from thawed leukapheresis product were isolated and activated with Dynabeads Human T-Activator CD3/CD28 magnetic beads (Invitrogen) and transduced with gammaretroviral 19-28z supernatant. Transduced T cells were then further expanded with the Wave bioreactor to achieve the desired modified T cell dose.

Assessment of 19-28z CAR-modified T cell persistence

Persistence of 19-28z CAR-modified T cells in patient peripheral blood and BM was assessed by FACS with biotinylated goat anti-mouse IgG F(ab′)2 (Jackson ImmunoResearch) as described (8). In the setting of low lymphocyte numbers in samples, persistence of CAR-modified T cells was assessed by FACS after nonspecific expansion of T cells ex vivo with Dynabeads ClinExVivo CD3/CD28 as previously described (8).

Assessment of anti-CD19 responses mediated by 19-28z CAR-modified T cells

Serial BM and peripheral blood samples were assessed for both detectable tumor (CD19+ CD10+) and normal B cells (CD19+ CD10) by flow cytometry. Persistence of normal B cells and the B-ALL clone was further monitored by deep sequencing (Adaptive Biotechnologies) (9). Briefly, 7.5 μg of genomic DNA (or its cell equivalent) from BM or blood was submitted to Adaptive Biotechnologies for multiplex PCR and resequencing. Sequences were then interrogated for copies of the malignant IgH clonotype, which was identified by multiplex PCR and resequencing of BM samples obtained during relapse. Additionally, MSK-ALL04 and MSK-ALL05 blood and BM were analyzed by FISH for the 9p21 deletion as an assessment of disease response to modified T cell therapy.

Analysis of cytokine profiles after 19-28z CAR-modified T cell infusion

Serial serum samples obtained before and after modified T cell infusion were analyzed for cytokine profiles as a surrogate marker for infused T cell activation with the Luminex IS100 system and commercially available 39-plex cytokine detection assays as described (8).

Chromium release assay

19-28z T cells were evaluated for CD19-targeted killing by a chromium release assay as described (13).

Clinical investigation statement

The Institutional Review Board at Memorial Sloan-Kettering Cancer Center (MSKCC) reviewed and approved this trial. All patients enrolled and treated on this trial gave written informed consent before participation. All clinical investigation was conducted according to the Declaration of Helsinki principles.

Supplementary Materials

Fig. S1. Trial schema for adult patients with relapsed B-ALL.

Fig. S2. Trial schema for adult patients with B-ALL in CR.

Fig. S3. Serum cytokine detection and correlation with tumor burden.

Fig. S4. Rapid hematopoietic recovery in MSK-ALL05 after infusion with 19-28z T cells.

Table S1. Characteristics of infused 19-28z CAR-transduced T cells; tumor burden in apheresis and BM.

Table S2. Adverse events.

Table S3. Percentage of 19-28z CAR+ T cells detected in the CD3+ T cells of the peripheral blood and of the BM (in parenthesis) of patients up to 57 days after CAR-modified T cell therapy.

References and Notes

  1. Acknowledgments: We thank V. Capacio, J. Hosey, and Y. Wang for excellent technical and quality control assistance. Funding: We thank the following for financial support: the National Cancer Institute (R.J.B., M.L.D., I.R., and M.S.), the American Society of Hematology–Amos Medical Faculty Development Program (M.L.D.), the Alliance for Cancer Gene Therapy (M.S.), the Carson Family Charitable Trust (R.J.B.), the William Lawrence and Blanche Hughes Foundation (R.J.B. and K.C.), the Mallah Foundation, the Majors Foundation and Mr. Lew Sanders (M.S., R.J.B., and I.R.), the Annual Terry Fox Run for Cancer Research (New York, NY) organized by the Canada Club of New York (R.J.B.), Kate’s Team, the CLL Global Research Foundation (R.J.B.), the St. Baldrick’s Foundation (K.C.), and Mr. William H. Goodwin and Mrs. Alice Goodwin and the Commonwealth Cancer Foundation for Research and the Experimental Therapeutics Center of MSKCC (M.S., R.J.B., and I.R.). Author contributions: M.S., R.J.B., and I.R. conceptualized the overall strategy and developed its clinical translation and implementation. The clinical protocol was written by M.L.D., M.S., and R.J.B. M.L.D. is the principal investigator of the protocol. Manufacturing of T cells, flow cytometry, and quantitative PCR acquisition of clinical samples was performed by C.T., J.S., J.Q., M.O., O.B.-O., Q.H., and T.W., supervised by X.W., and directed by I.R. Data from manufacturing and FACS/quantitative PCR monitoring were analyzed by X.W. and I.R. The manuscript was written by R.J.B., M.S., M.L.D., I.R., and J.P. R.J.B., M.S., I.R., J.P., and X.W. discussed and interpreted the results. L.G.C. performed statistical analyses and analyzed correlation between cytokine levels and tumor burden. R.J.B., M.L.D., J.P., R.K., K.C., P.S., J.J., T.R., and M.F. enrolled patients to the protocol and/or managed the leukemia patients in Memorial Hospital. I.V.R. and C.H. designed and performed molecular assays to identify the malignant IgH clonotype associated with the leukemia cells of enrolled and treated patients. P.M. evaluated all pre- and posttreatment bone marrow aspirates for evidence of leukemia. Y.B. is the Research Study Assistant for the protocol and assisted with enrollment, sample acquisition, and data safety monitoring of patients. Competing interests: M.S. and R.J.B. are co-holders of U.S. Patent 7,446,190, which covers the 19-28z receptor. The other authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article