Targeting CD38 in T-ALL Upregulates the Targetable Polyamine Metabolism Pathway

Introduction: Relapsed or primary refractory T-cell acute lymphoblastic leukemia (T-ALL) is associated with very poor salvage rates. CD38 is a ubiquitous cell surface protein in pediatric T-ALL, expressed in >95% of cases (PMID: 32439800). Previous work has shown the CD38 targeting monoclonal antibody Daratumumab (Dara) is efficacious in patient derived xenograft (PDX) models of T-ALL in NSG mice (PMID: 29305553) and is currently being investigated in clinical trials in pediatric T-ALL (NCT03384654). Additionally, chimeric antigen-receptor T-cells targeting CD38 (CART38) in development show efficacy in T-ALL and other hematologic malignancies in preclinical models (Glisovic-Aplenc et al. In Review). CD38 is a type II-transmembrane glycoprotein that has been implicated in the regulation of cytoplasmic calcium flux. The mechanism by which targeting CD38 in T-ALL induces cell death is not known. Given the efficacy of Dara in an immunocompromised mouse model, we hypothesized that binding of Dara to CD38 on T-ALL blasts may induce cell death by a direct mechanism.

Methods: We used CRISPR/Cas9 to knock out (KO) CD38 from blasts in the T-ALL cell lines HPB and MOLT-4 and from the B-ALL cell line NALM-6. CD38KO blasts were compared to those treated with non-targeting RNA and to unmanipulated blasts. In animal experiments, PDX mice were injected with T-ALL blasts modified to express luciferase to evaluate different therapies alone and in combination, with short and long-term exposure. Non-invasive bioluminescent imaging was used to measure radiance as a surrogate marker of disease burden. We screened our bank of 75 T-ALL xenografts for an endogenously CD38 negative PDX and identified one PDX (TH67).

RNA sequencing was performed using Illumina NextSeq 500 and llumina RNA True Seq Access for library preparation. Ultra high-performance liquid chromatography/tandem accurate mass spectrometry (UHPLC/MS/MS) was performed by Metabolon Inc. to measure 660 known metabolites and identify perturbed metabolic pathways.

Results: We assessed the role of CD38 in T-ALL by investigating: (1) the impact of two different CD38-targeted immune therapies (Dara and CART38) in T-ALL PDX models; (2) the impact of CD38KO on T-ALL cell lines; and (3) the differences between T-ALL blasts that did/did not express CD38 at baseline. First, we performed RNAseq to assess the impact of disruption of CD38 in T-ALL blasts. Using KEGG, we found the top differentially expressed genes after CD38 KO were those involved in cell metabolism and lysosomal processing. Given the putative role of CD38 in cellular metabolism in normal hematopoietic cells and our gene expression data, we subsequently focused on evaluating changes in cellular metabolism in T-ALL blasts using UHPLC/MS/MS.

As in normal hematopoietic cells, we found disruption of CD38 led to aberrant nucleotide metabolism with elevated ATP, ADP, and adenosine 2’3’ cyclic monophosphate. We made the novel observation that polyamine metabolism was significantly impacted by targeting CD38, with increased levels of polyamines (including spermine and spermidine) after CD38KOs. Further, we found increased levels of the same polyamines in PDX samples treated with Dara or CART38. The endogenously CD38(-) PDX TH67 also had significant elevations of spermine and spermidine relative to unmanipulated TH34 blasts. The degree of change varied based on treatment with CART38 or Dara and with CD38KO, as well as in different cell lines and PDX models but with a consistent pattern. To test the hypothesis that blockade of the polyamine pathway could potentiate CD38 targeting therapies, we used the ornithine decarboxylase inhibitor difluromethylornithine (DFMO) in combination with Dara. Five mice in each arm were randomly assigned to received Dara alone, DFMO alone, saline or combined Dara/DFMO. Dara/DFMO was well-tolerated and outperformed either agent alone.

Conclusions: Using metabolic profiling we identify novel changes in the polyamine synthesis pathway in CD38KO cell lines and in blasts treated with Dara and CART38. Blockade of the polyamine synthesis pathway with DFMO potentiates the action of dara in a PDX model of pediatric T-ALL. Ongoing studies are combining DFMO with CART38 in PDX. DFMO is a widely available, FDA-approved medication that could be easily translated into clinical trials.