Immunotherapy in Pediatric Acute Lymphoblastic Leukemia

Approximately 85% of ALL cases are B-ALL [2]. Cure rates for B-ALL significantly rose over the past five decades from 10% to 90% [1-3] due to multi-agent chemotherapy regiments, CNS prophylaxis and better risk stratification [3]. Despite these successes, about 2% of patients are refractory to chemotherapy and another 10% to 15% of patients will relapse [4]. Treatment for these patients remains a therapeutic challenge. Event free survival for Abstract


T-ALL
T-ALL accounts for approximately 15% of pediatric ALL cases [1,10]; and historically these patients have inferior outcomes to patients with B-ALL with event-free and overall survival around 70% and 80% respectively [11][12][13]. Survival has improved with intensification of therapy and T-cell focused regimens, such as the addition of nelarabine to treatment paradigms [14]. However, survival after relapse is about 30% due to a lack of effective salvage therapies [15].

Immunotherapy
Immunotherapy is a revolutionary treatment aimed to improve survival and reduce the toxicity of chemotherapy by harnessing the patient's own immune system to target cancer cells. Several different approaches have been developed. Antibody therapy utilizes antigens present on the surface of leukemia cell to aid in the immune system's attack of the cancer cell. Therapies include monoclonal antibodies, antibody drug conjugates and Bispecific T-cell engagers (BiTES). Adaptive therapies manipulate patient's cytolytic immune cells to recognize tumor cells and elicit an anti-tumor response. These therapies include chimeric antigen receptor T-cell (CAR-T) therapy. This review will focus on immunotherapeutic options approved and under investigation for pediatric ALL. Common targets are highlighted in Tables 1 and 2

Inotuzumab ozogamicin (InO)
Mechanism of action: CD22 is expressed on 80% to 90% of B-ALL cells. Inotuzumab ozogamicin (InO) is a humanized anti-CD22 monoclonal antibody conjugated to the cytotoxic drug, calicheamicin [16]. Calichaemicin is cleaved and binds to minor DNA grooves causing doublestranded DNA breaks and apoptosis of the leukemia cell [16]. Clinical trials of InO in B-ALL are highlighted in Table 3.
Adult experience with InO: U.S. Food and Drug Administration (FDA) approval of InO for relapsed/ refractory CD22 positive ALL was based on the INO-VATE trial. The study showed a superiority of InO compared to standard of care (SOC) chemotherapy with improved complete remission (CR)/remission with partial hematologic recovery (CRh) rates, 73.8% vs. 35% and progression-free survival (PFS) 5.0 months vs. 1.7 months. More patients proceed to transplant in the InO arm (48% vs. 22%, p<0.0001) [17]. At the two-year follow-up, overall survival (OS) rates were superior with InO 22.8% and 10.0% [17].
Pediatric experience with InO: Of five pediatric patients with relapsed CD22 positive B-ALL treated with InO as part of an adult phase 2 trial, three had a CR/CRh [18]. A retrospective analysis of compassionate use of InO in fifty-one pediatric patients with relapsed/refractory B-ALL showed a 12-month EFS and OS of 23% and 36% respectively [19]. Twenty-one patients underwent a hematopoietic stem cell transplant (HSCT) after achieving CR [19]. In the Children's Oncology Group (COG) trial AALL1621 (NCT02981628) of InO in heavily pre-treated relapsed/refractory, CD22 positive B-ALL patients, 58.3% had a CR/CRh, with 65.4% of those having a minimal residual disease (MRD) response [20].
Ongoing Trials with InO in pediatrics: There are several ongoing pediatric trials investigating the timing and indications for InO. InO is being studied in upfront therapy, relapsed/refractory MRD positive ALL with chemotherapy, and as consolidation post-transplant (Table 4).

Sinusoidal
obstructive syndrome (SOS): Sinusoidal obstructive syndrome (SOS) was seen more commonly in patients who are treated with InO than salvage chemotherapy, a serious concern in patients whom subsequent transplant is a consideration. In INO-VATE trial, rates of SOS were 14.0% (5% fatal) in the InO arm vs. 2.1% in the SOC chemotherapy arm [17]. Risk factors for SOS included conditioning with dual alkylators, hyperbilirubinemia before HSCT, and prior HSCT (OR 6.02; p=0.032) [17] [20] at 11 of the 21 (52%) patients in compassionate use study [19]. Strategies to prevent SOS include avoiding dual alkylating agents and/or thiotepa and hepatoxic agents, prophylactic ursodiol, proceeding to HCT after two cycles of InO, and close monitoring for SOS [21]. Ongoing trials include Levocarnitine and Vitamin B to reduce hyperbilirubinemia with InO treatment (NCT03564678), lowering the dose of InO pre-transplant (NCT03677596) and reduced intensity transplant (NCT03856216).

Daratumumab
Mechanism of action: CD38 is a type II transmembrane glycoprotein on the surface of thymocytes, activated T-cells and terminally differentiated B cells, with low level expression on other normal lymphoid and myeloid cells [22]. CD38 expression has been seen on T-ALL blasts and remains stable after treatment with chemotherapy [22]. Daratumumab is a human monoclonal antibody directed against CD38 [22]. It is FDA approved for multiple myeloma both as monotherapy and in combination [23,24]. Preclinical data has shown efficacy of Daratumumab in T-ALL models [22,25], and case series have shown efficacy as salvage therapy in relapsed T-ALL [26,27]. There are ongoing clinical trials of Daratumumab in pediatric T-ALL and B-ALL in combination with cytotoxic chemotherapy (NCT03384654). Other monoclonal antibodies being tested in T-ALL are listed in Table 2.

BiTEs
Bispecific T-cell-Engaging (BiTE) antibodies are antibodybased molecules that bind to distinct surface markers on T-cells and tumor cells to form the immunological synapse [28,29]. BiTEs bind the invariant signaling component of the T-cell receptor (TCR), CD3, and a surface target antigen on tumor cells, resulting in T-cell activation, expansion and tumor cell lysis [28,29]. BiTEs are independent of T-cell receptor specificity and do not require MHC presentation of the antigen; thus, bypassing T-cell regulation [29]. Unlike CARTs, BiTEs do not require manufacturing and infusion of T-cells [29].

Blinatumomab
Mechanism of action: CD19 is expressed on approximately 90% of B-ALL cells [30]. Blinatumomab is a BiTE that binds to CD19 on leukemic cells and CD3subunit of the TCR on T-cells [29]. Clinical trials of Blinatumomab for B-ALL are highlighted in Table 5.

Role in relapsed/refractory B-ALL:
In 2014, the FDA granted accelerated approval of blinatumomab for adult Philadelphia chromosome negative (Ph-) relapsed/ refractory B-ALL based on a single-arm study of 189 adults that showed efficacy and manageable toxicity [31]. Eightyone patients (43%) had a CR/CRh within two cycles of blinatumomab [31]. Median overall survival was 6.1 months [31]. This was superior to historical controls who received SOC, salvage chemotherapy [31]. Efficacy was confirmed in the TOWER trial, a multicentered, randomized, phase III trial comparing blinatumomab to chemotherapy in adult relapsed/refractory Ph-ALL [32]. CR was achieved in 91 patients (34%) in the blinatumomab arm compared to 21 patients (16%) in the SOC arm. The median overall survival was significantly longer for the blinatumomab arm (7.7 months versus (4.0 months) in the SOC arm [32].

Role in Philadelphia chromosome-positive (Ph+) ALL:
The approval of blinatumomab was extend to Ph+ relapsed/refractory B-ALL based on the ALCANTARA trial showing a 36% CR/CRh, with 88% complete MRD response in patients with relapsed/refractory Ph+ ALL, previously treated with TKI treatment [33]. Blinatumomab as consolidation to treatment with TKI has also been studied in a multicenter phase II trial of Ph+ ALL, patients  were treated with dasatinib, followed by post-induction consolidation with blinatumomab. At the end of two cycles of blinatumomab 19/35 (54%) had a molecular response that further increased after subsequent cycles [34]. Twelvemonth OS and DFS are 96.2% and 91.6% respectively [34].
There are several ongoing studies examining the efficacy of TKIs with blinatumomab (Table 6).

Role in MRD positive disease:
In 2018 the FDA granted approval for blinatumomab for the treatment of adults and children with B-ALL in a morphological first or second CR with MRD [35]. Eighty-eight of 113 patients (78%) achieved a complete MRD response after one cycle of blinatumomab [35]. Patients who achieved a complete MRD response had a prolonged OS (38.9 vs 12.5 months; p=0.002) and RFS (23.6 vs 5.7 months; p=0.002) [35].
Role in first vs. later relapse: Blinatumomab appears to be a more effective salvage therapy in first versus second or later relapse. In the TOWER study, blinatumomab's effect on overall survival was greater for first salvage therapy (HR 0.59; p=0.016) than second or greater salvage therapy (HR 0.72; p=0.055) [36]. Similarly, in the BLAST MRD trial, patients who had previously relapsed had inferior RFS and OS compared with those treated in first remission ( [38]. Data for the low risk (LR) randomization is pending. There are ongoing studies investigating the role of blinatumomab in upfront therapy including the COG trial AALL1731 (NCT03914625) that is studying the addition of blinatumomab to standard chemotherapy in patients with NCI SR B-ALL at high risk for relapse. Blinatumomab is also being studied in HR/IR newly diagnosed B-ALL in the European Studies AIEOP-BFM ALL 2017 (NCT03643276) and PETHEMA-BLIN-01 (NCT03523429). Lastly, blinatumomab is also being studied as maintenance after allogenic HSCT (NCT02807883 & NCT03114865) ( Table 6). Combining blinatumomab with other immunotherapies is also being investigated. There is an ongoing adult trial combining treatment with inotuzumab ozogamicin with mini-HCVD with or without blinatumomab in previously untreated acute lymphoblastic leukemia, (NCT01371630). In AYA patients, blinatumomab and inotuzumab ozogamicin are being studied in newly diagnosed and relapsed/refractory CD22+ B-ALL (NCT03739814). The ability of checkpoint inhibitors to further enhance the efficacy of blinatumomab is also actively being studied (NCT03605589, NCT03512405, NCT03160079, NCT02879695).
Biomarkers to predict response: Predictive biomarkers of response to blinatumomab are emerging.
Patients who have a lower baseline disease burden [31] and day 15 MRD have a better response [39]. In addition, superior response was correlated with greater T-cell expansion of effector memory T-cells [40] and a higher percentage of regulatory T-cells [41]. Identifying additional biomarkers to determine response is actively being studied.

CAR-T Therapy
Chimeric antigen receptors (CARs) are T-cells that are engineered to recognize tumor associated antigens. CARs are composed of T-cell signaling moiety and a tumor specific antigen binding domain, commonly a single-chain variable-fragment monoclonal antibody that is fused to a transmembrane domain [42]. Various generations of CARs have been developed to heighten function based on the knowledge that T-cells require two signals to be activated, T-cell receptor (TCR) engagement and co-stimulation. First generation CARs consisted of T-cell receptor complex domain and antigen recognition domains, only providing signal 1; whereas, second generation CARs were constructed to contain co-stimulatory signaling domains including CD28, 4-1BB (CD137), and OX40 (CD134) [42]. Third generation CARs further enhanced T-cell signaling by containing tandem cytoplasmic signaling from two co-stimulator receptors (CD28-4-1BB or CD28-OX40) [42]. Fourth generation CARs have pro-proliferative T-cell costimulatory ligands (4-1BBL) or proinflammatory cytokines (IL-12) [42]. Advantages of CAR-T therapy include HLA-independent recognition of tumor antigen; allowing T-cells to recognize the antigen as foreign and activity is unaffected by HLA down regulation in tumor cells. In addition, both CD4+ and CD8+ T-cell subsets are transduced, allowing for both helper and cytotoxic activity.

CD19 CAR-T
Commercial approval: There are two FDA approved CD19 directed CAR-T products, tisagenlecleucel (CTL019) and axicabtagene ciloleucel. Both are second generation CAR-T-cells. Tisagenlecleucel uses a 41BB costimulatory domain and is transduced by lentivirus, whereas axicabtagene ciloleucel uses the CD28 costimulatory domain and is produced by retroviral transduction. Tisagenlecleucel is FDA approved for relapsed/refractory B-ALL in pediatric and young adult patients. Axicabtagene ciloleucel is approved for relapsed/refractory B-cell lymphoma in adults and is being studied for the treatment of pediatric B-ALL.
The FDA approval of tisagenlecleucel was based on a pivotal, global multicenter trial of tisagenlecleucel in pediatric relapsed/refractory, CD19+, B-ALL that showed an overall remission rate of 81%, all with MRD response [43]. Six-and 12-month relapse-free survival rates were 80% and 59% respectively [43].
Clinical trials: Trials have shown second generation CD19-CAR-T therapy induced remission in heavily pre-treated patients with multiple relapsed/refractory B-ALL [43][44][45][46]. Axicabtagene ciloleucel is currently being studied in pediatric patients with relapsed/refractory B-ALL previously treated with salvage therapy or HSCT (NCT02625480). In adult relapsed/refractory B-ALL, 44 of the 53 patients (83%) had a CR and 32 patients (67%) had a MRD response [44]. In a Phase 1 trial of 21 pediatric patients with relapsed/refractory B-ALL or Non-Hodgkin's Lymphoma, CR was seen in 66.7% (14 of 21) of patients with a MRD response occurring in 60% of patients [45]. CD19-CAR-T therapy is being studied in newly diagnosed very high-risk (VHR) B-ALL patients in the COG trial AALL1721 (NCT03876769) and St Jude Total Therapy XVII (NCT03117751). In addition, CD19-CAR-T is being studied in combination with checkpoint inhibitors to enhance efficacy for the CAR-Ts and decrease T-cell exhaustion (Table 8).

Unanswered questions:
Several unanswered questions remain including the role of CD19-CAR-T in upfront therapy and if it should be used as monotherapy versus a bridge to HSCT. Historically, patients at high risk of relapse, myeloablative transplant is recommended. In the Park et al. study, subsequent transplant did not influence EFS or OS for the patients who had a MRD response after CD19-CAR-T [44]. In the Eliana study, the 18-month EFS was 66% with a median persistence of CAR-T of 168 days [43,47]. Eight patients underwent allogeneic hematopoietic stem-cell transplantation while in remission [43,47]. Conversely, 29 (45%) patients had an ongoing response without additional treatment, and 19 patients (29%) relapsed without receiving additional therapy [43,47]. An association has been seen between early loss of B-cell aplasia with relapse. In patients with early loss of B-cell aplasia, if there is an available donor and the patient is in good functional status, early transplant is recommended. Long term follow-up is needed to better assess which patient's CAR-T can be used as monotherapy.

CD22 CAR-T
Studies of CD22 showed a similar anti-leukemic effect and safety profile to CD19-CARs. In a phase I trial of CD22 BB.z CART in heavily pretreated relapsed/refractory patients, 12/21 (57%) of patients had a CR, with nine patients having a MRD response (NCT02315612) [48]. To further potentate the efficacy of CD22, CAR-T Bryostatin 1 has been seen to upregulate CD22 on leukemia cell lines and improve CART function and persistence [49].

Dual targeting CAR-T
Antigen-escape relapse after CD19 directed therapies is a major challenge, thus dual targeting of CD19 and CD22 is being developed. Bi-cistronic CAR-T that express CD19 and CD22 scFv simultaneously on every cell and mono-CARs that express CD19 and CD22 scFv separately have been developed as dual-target CAR-T cells. Phase 1 trials of different dual targeting CAR-T therapies are highlighted in Table 8

CAR-T for T-ALL
Several challenges have been encountered in the development of CAR-T including disease heterogeneity, T-cell aplasia, fratricide, and increased side effects in T-ALL. There is a large amount of disease heterogeneity in T-ALL due to distinct stages when T-cell differentiation arrest occurs, making identifying a target difficult [50]. Furthermore, targets on T-lymphoblasts are likely to be on normal T-cells leading to a severe immunocompromised state and fratricide of CAR-T. Fratricide of the CAR-T product or the destruction of the CAR-T due to the target being on both the malignant T-cells and on the CAR-T, leads to decreased CAR-T expansion and persistence. CD3 and CD7 CAR-Ts are more prone to fratricide compared to other targets such as CD1a and CD5 [51,52]. Gene editing to decrease expression of the target antigen on the CAR-T is being studied [53,54] in addition to "off-the-shelf" CAR-T without the target antigen [55]. Preclinical efficacy has been shown in NK-CARs [56,57].

Challenges in CAR-T manufacturing
One of the challenges of CAR-T is manufacturing the product. For adequate collection of T-cells, it requires an absolute lymphocyte count ≥ 500 cells/µL or an absolute CD3 count ≥ 150 cells/µL. This is particularly challenging in heavily pre-treated patients due to poor bone marrow in patients with a higher cumulative dose of chemotherapy [58] and in younger patients due to their size. In the phase II study of tisagenlecleucel, eight patients did not receive the CAR-T infusion due to manufacturing related issues and another seven died before infusion [43,47]. Early collection is suggested for high risk patients; and gene-edited, universal CAR-T-cells are in development. Allogeneic CD19-CAR-T-cells successfully treated two 177 NCT03448393 CD19/CD22 CAR-T in R/R B-ALL  infants with B-ALL using non-HLA-matched, universal, CAR19 (UCART19) T-cells manufactured from a healthy female donor [59].

Persistence of CAR
There have been several mechanisms proposed explaining why certain patients do not respond or have a durable remission following treatment with CD19-CAR-T-cells. One mechanism of relapse is poor persistence of the CAR-T-cell [46,60]; however, the length of CAR persistence required to induce a durable response or cure is unknown. There is no commercially available test to detect CAR-T persistence. B-cell aplasia has been used as a marker with early emergence of CD19-positive B-cells, within six months of CAR-T infusion, being associated with early relapse [46,60]. The presence of hematogones in the bone marrow has also been suggested as an earlier marker of loss of persistence and can occur while B-cell aplasia is still present [61]. Persistence may also be influenced by the CAR-T construct. The 19-BBz CAR are more persistent (168 days) than 19-28zCAR (~28 days), and 19-BBz CAR-T are associated with longer remission without HSCT [43,45]. Another contributor to decreased persistence is the development of T-cell mediated anti-CAR immune response related to the murine CD19 scFV [45]. Re-infusion with humanized anti-CD19 CAR T-cells has induced remissions in children and young adults with relapsed/refractory B-ALL previously treated with murine-CD19-CAR-T [62]. Lastly, expansion and persistence are improved with CAR-T generated from early linage T-cells (naïve T-cells and stem central memory T-cells) versus more differentiated T-cells (effector memory and terminal effector cells) [63,64]. Similar to studies in blinatumomab, T-cell exhaustion and high levels of T-regulatory cells have also been thought to contribute to treatment failure due to poor CAR-T persistence [65,66]. Most CAR-T protocols include lymphodepletion prior to CAR-T infusion, which leads to depletion of regulatory T-cells and greater engraftment. In addition, it is felt checkpoint inhibition may mitigate T-cell exhaustion. Re-expansion of CAR-T-cells has been seen in patients who are treated PD-1 inhibitors after early loss of CAR-T-cells or relapse [67].

Antigen escape
In antigen directed therapy, escape, or loss of the therapy directed antigen on tumor cells is commonly seen in relapse [37,43]. Mechanisms of CD19 escape seen with blinatumomab and CD19-CAR-T include: selecting for pre-existing antigen negative leukemia, trogocytosis, the development of mutations or alternate splice variants of CD19, or lineage switching [68][69][70][71][72][73][74]. There are ongoing trials using two immunotherapies targeting different antigens and bispecific CAR-T-cells (Table 8) to mitigate this effect.

Side effects of immunotherapy
Cytokine release syndrome (CRS): Cytokine release syndrome (CRS) is a systemic inflammatory response due to a rise in cytokine levels during T-cell activation and expansion. Symptoms range from mild and self-limiting to severe and life-threatening and consists of fever, myalgia, capillary leak, hemodynamic instability, coagulopathy and multi-organ failure [75]. Higher disease burden has been associated with higher grade CRS [60]. Varying grading symptoms have been developed, and ASBMT consensus grading system was developed last year [76]. Tocilizumab, an IL-6 receptor antagonist, has been shown to be effective in treating CRS [76]. Other medications that have been considered include infliximab, etanercept, and anakinra [75]. There are ongoing studies regarding the optimal timing of Tocilizumab administration where patients with a higher disease burden will receive early Tocilizumab (NCT0290637).
Neurotoxicity: Neurotoxicity has also been seen with immunotherapy. Symptoms include delirium, encephalopathy, aphasia, lethargy, seizures and cerebral edema [76]. Symptoms typically occur either during or more commonly after CRS. The ASBMT similarly recently created an Immune effector-cell associated encephalopathy (ICE) score [76]. Corticosteroids are recommended for severe neurotoxicity.
B-cell aplasia: B-cell aplasia is an on-target, off tumor adverse effect of immunotherapy directed to antigens on normal B-cells including CD19, CD20 and CD22. B-cell aplasia occurred in all patients who responded to Tisagenlecleucel, and 83% experienced B-cell aplasia for at least 6 months [43,47]. Immunoglobulin replacement is recommended following treatment while there are signs of B-cell aplasia.

Future Directions
The development of immunotherapy is a major advancement in treating pediatric ALL. Particularly in relapsed and refractory B-ALL, immunotherapy has been able to induce remission in chemotherapy refractory patients who had limited treatment options. The timing of immunotherapy in treatment paradigms is being investigated including the role in upfront, salvage, consolidation, and maintenance therapy. Furthermore, checkpoint inhibitors are being studied to further enhance the efficacy of many immunotherapies (Table 9). The role of immunotherapy in T-ALL has remained a challenge, and further research into optimal targets to limit effects on normal T-cells and maximize the efficacy of the therapy is ongoing. Preclinical and clinical research has shown significant promise in improving survival for these patients.