Work

Phenotype switch in acute lymphoblastic leukaemia associated with 3 years of persistent CAR T cell directed‐ CD 19 selective pressure

Year: 2019

Type: article

Abstract: Genetically modified T cells that express a CD19-specific chimeric antigen receptor (CAR T cells) have shown impressive and durable efficacy in the treatment of chemo-refractory and relapsed acute lymphoblastic leukaemia (ALL) (Maude et al, 2018; Park et al, 2018). Although most relapses in the setting of CD19-directed immunotherapy occur due to alternative splicing resulting in epitope loss (Sotillo et al, 2015), a rare mechanism of immune escape characterized by acquisition of a CD19-negative myeloid lineage switch has also been reported (Gardner et al, 2016). In these patients, a rapid emergence of myeloid cells with similar cytogenetic abnormalities, including a persistent KMT2A rearrangement, led the authors to conclude this represented a true lineage switch rather than a biphenotypic leukaemia at diagnosis. This rapid emergence of a myeloid clone has also been observed in patients following blinatumomab therapy, a bispecific antibody targeting CD19 (Haddox et al, 2017). We report a case of a paediatric patient with B-cell ALL (B-ALL) who experienced a rare lineage switch to a myeloid phenotype 3 years after infusion of SCRI-CAR1v1, a CD19-specific CAR T cell product. This male patient initially presented at 5 years of age with B-ALL diagnosed by the leukaemic expression of CD19, CD20, CD34, variable TdT and weak variable CD2 (Table 1). He completed 3·5 years of standard paediatric therapy without major complications, but experienced a combined medullary and central nervous system recurrence 5 months after completion of therapy (Table 1). At this recurrence, cytogenetics showed a possible clonal deletion of the 3′portion of the KMT2A locus (61%) not previously seen in his initial diagnosis. He was able to achieve a complete remission (CR) with negative minimal residual disease (MRD) by flow cytometry following several re-induction regimens. He then proceeded to a matched unrelated donor bone marrow transplant in second CR. Approximately 5 years after transplant, he experienced another medullary recurrence refractory to chemotherapeutic therapy. He was enrolled on the PLAT02 study (NCT02028455), received SCRI-CAR19v1 therapy complicated by mild cytokine release syndrome (CRS), then achieved a CR with negative MRD. He continued to have functional persistence of the CAR T cells with no detectable circulating B cells and remained on immunoglobulin support for 3 years. Approximately 3 years post-SCRI-CAR19v1 infusion, he developed a cutaneous mass over the left scapula as well as a large testicular lesion that were excised and identified as myeloid tumours expressing CD43, CD45, CD68, CD163 (weak) and lysozyme. Further molecular testing revealed a t(11;11)(q23.3;q23.3) KMT2A-USP2 fusion as well as PTPN11 D61V and PHF6 T12fs*23 mutations that were present in his B-ALL sample before SCRI-CAR19v1, confirming the same clonal lineage (Table 1). The PTPN11 D61V mutation is a known proliferative driver leading to RAS-MAPK activation. To assess whether the PTPN11 mutation played a role in tumour proliferation, the leukaemia cutis sample was assessed for phosphorylation of MAPK3/MAPK1 (Fig 1). Haematoxylin and eosin staining showed infiltration of the dermis with leukaemic cells consisting of an atypical proliferation of large, oval to spindled cells with convoluted nuclei, vacuolated cytoplasm and interspersed mitoses (Fig 1). These abnormal cells, expressing CD45, also stain for anti-phospho-MAPK3/MAPK1, similar to a cutaneous melanoma metastatic sample with a BRAF V600E mutation known to stimulate the MAPK pathway. While lineage conversion from a lymphoid to a myeloid phenotype has been described as a rare event in patients on standard therapy (Rossi et al, 2012), usually in the presence of KMT2A-rearranged acute leukaemia, it appears to be increasing in patients following treatment with CD19-directed therapy. This case is unique when compared to cases of previously reported lineage switch in that the time to relapse was much later and was restricted to extramedullary recurrence. In this case, it is unclear temporally when the lineage switch occurred, but it is remarkable that the period of latency was 3 years before the recurrence appeared. Prior reported lineage switches following selective CD19 pressure have tended to occur within a month of the treatment, with the latest reported case occurring at around 1 year (Jacoby et al, 2016; Haddox et al, 2017). PTPN11 is a ubiquitous protein tyrosine phosphatase essential in many signalling pathways. PTPN11 function is required for normal activation of the RAS-MAPK pathway and its downstream transcriptional targets for proliferation. Gain-of-function mutations of PTPN11, such as D61V, are seen in ~5–10% of acute myeloid leukaemia (AML) (Hou et al, 2008). Currently, single agent therapies that target the RAS-MAPK pathways have garnered much attention due to exciting activity of these agents in certain cancer subsets. Further studies have described a strong interaction of germline PTPN11 mutations within the environment driving protection of leukaemic stem cells (Dong et al, 2016). Our patient exhibited a pathogenic activating mutation in PTPN11 within the leukaemic cells, suggesting a different mechanism by which these stem cells remained protected and dormant for over 3 years. Lastly, this patient's tumour also carried a probable pathogenic truncating mutation in the transcription factor PHF6. Loss of function mutations in PHF6 have recently been implicated in leukemogenesis and lineage plasticity (Soto-Feliciano et al, 2017). Although previously described in AML and T-ALL, there are no prior studies that identify this mutation in B-ALL. Most intriguing, Soto-Feliciano et al (2017) described a mouse model of phenotype plasticity in leukaemic cells from PHF6-knockout mice induced by the oncogene BCR-ABL1, further illustrating the importance of this transcription factor in navigating selection pressure. It remains unknown as to whether mutations in KMT2A, PTPN11 and PHF6 enable leukaemic cells to continue to proliferate and develop phenotypic plasticity to escape continued pressure against CD19 expression after 3 years of continued exposure. One might posit that the KMT2A and loss of PHF6 led to aberrant chromatin remodelling, enabling phenotype plasticity with PTPN11 activity driving continued proliferation. Ultimately, strategies aimed at combating immune escape require an understanding of the molecular underpinnings that enable resistance. Collectively, this case illustrates the presence of prolonged selection pressures for years and raises questions about the potential utility of combined targeted and immunotherapy in high-risk patients with actionable mutations. We wish to thank the OHSU Histopathology Core services for the processing of the tissue. OML and KP analysed the data and wrote the manuscript, TF obtained the tissue samples and contributed to the writing of the manuscript, JD analysed the immunophenotype and contributed to the writing of the manuscript, RP contributed essential next-generation sequencing, analysed the data and contributed to the writing of the manuscript, RAG and BHC designed the study, analysed the data and contributed to the writing of the manuscript. OML, PK, TF, JD, RP, RAG and BHC have no competing financial interests.

Source: British Journal of Haematology

DOI: 10.1111/bjh.15812