AUTHORS
Zhiwei Ang, PhD, Annette Castro2, Luca Paruzzo, MD3, Carolin Schmidt4, Tori Fuller5, Alberto Carturan, MD6, Katharina Hayer7, Christopher Kwok8, Jacinta Davis, PhD9, Samantha Soldan10, Federico Stella, MD11, Manuel Torres-Diz, PhD12, Rawan Shraim, PhD13, Patricia King Sainos14, Kayla Ji15, Patrick Krohl16, Justyn Fine17, Shannon McGettigan18, Priyanka Sehgal19, Daniel Martinez20, Patrizia Porazzi, PhD21, Jamie B Spangler, PhD22, Paul Lieberman23, James L Riley24, Vinodh Pillai, MD25, Haley Newman, MD, MSCE26, Martin Carroll, MD27, Zainul Hasanali, MD, PhD28, David Allman, PhD29, Dan T. Vogl, MD, MSCE30, Caroline Diorio, MD31, David T. Teachey, MD32, Marco Ruella, MD33, Andrei Thomas-Tikhonenko, PhD34
AFFILIATIONS
1 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
2 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
3 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, United States
4 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
5 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
6 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, United States
7 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
8 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
9 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
10 The Wistar Institute, Philadelphia, United States
11 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, United States
12 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
13 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
14 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
15 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
16 Johns Hopkins University, Baltimore, United States
17 Johns Hopkins University, Baltimore, United States
18 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, United States
19 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
20 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
21 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, United States
22 Johns Hopkins University, Baltimore, United States
23 The Wistar Institute, Philadelphia, United States
24 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, United States
25 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, United States
26 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
27 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, United States
28 Indiana University, Bloomington, United States
29 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, United States
30 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, United States
31 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
32 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
33 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, United States
34 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
ABSTRACT
Chimeric antigen receptor (CAR) T-cell therapies are among the most effective treatments for relapsed/refractory B-cell acute lymphoblastic leukemia (B-ALL), B-cell lymphoma (BCL) or multiple myeloma (MM). However, these CD19- or BCMA-directed therapies are not curative for large subsets of patients with antigen-negative relapses. Meanwhile, designing CAR T-cells for T-cell acute lymphoblastic leukemia (T-ALL) is hindered by a lack of suitable antigens to discriminate between T-ALL and normal T cells. Targeting T-cell lineage markers causes T-cell aplasia and CAR T-cell fratricide, requiring additional genome-editing steps. Hence, there is a clinical need for novel therapeutic strategies that avoid fratricide and circumvent antigen escape in T-ALL, B-ALL, BCL and MM.
By mining bulk and single-cell RNA-sequencing datasets, we determined that the purinergic receptor P2X5 (P2RX5) has the same B-cell lineage-restricted expression profile in normal postnatal tissues as CD19. Notably, unlike CD19, P2RX5 is expressed not only in B-ALL and BCL, but also in T-ALL and MM. This makes P2RX5 an attractive pan-lymphoid malignancy target for CAR T-cell therapies. Although the surface localization of P2RX5 as an ATP-gated ion channel should make it readily accessible to CAR T cells, human P2RX5 is widely regarded as a pseudogene. This is due to two pervasive protein-truncating variants, namely the rs891746 single nucleotide polymorphism that causes skipping of exon 10 and the rs3215407 indel that introduces a premature STOP codon. However, this pseudogenization is not universal, and up to 80% of individuals of African ancestry retain at least one copy of the evolutionarily conserved full-length P2RX5 allele (P2RX5-FL). Using overexpression and CRISPR/Cas9-mediated editing, we have shown that while truncated P2RX5 protein variants are rapidly degraded, P2RX5-FL is properly expressed on the cell surface.
Here, we present mechanistic and preclinical data on P2RX5-FL as a novel CAR T-cell target. The persistent induction of P2RX5 expression in T-ALL, but not in mature T cells, may reflect the immature thymic origin of T-ALL. Indeed, while absent in newborn and adult thymocytes, P2RX5 mRNA is robustly expressed in double-negative thymocytes of the prenatal human thymus. Alternatively, P2RX5 expression could be driven by oncogenic Notch, as treatment with small-molecule Notch inhibitors reduced P2RX5 expression across all T-ALL cell lines tested.
To characterize P2RX5 expression at the protein level, we generated a P2RX5-reactive mouse monoclonal antibody (clone 1D9). When tested for cross-reactivity against >7,000 native proteins in the Integral Molecular Membrane Proteome Array, 1D9 showed minor off-target binding only to Nectin-4, a tumor-associated antigen with restricted expression in normal tissues. Importantly, this cross-binding was only detectable in cancer cell lines with Nectin-4 overexpression, with undetectable binding to Nectin-4-positive normal cells, such as primary keratinocytes. Tissue staining with 1D9 closely tracked with P2RX5 mRNA profiles, with homogeneous staining of both live and formalin-fixed paraffin-embedded BCL, T-ALL and MM samples. In normal tissues, 1D9 staining was restricted to B-cell enriched organs such as peripheral blood, spleen and lymph nodes, with no detectable staining in the skeletal muscle, heart, liver, kidney, small intestine, colon, lung and brain.
Next, we designed a second-generation CAR containing the 4-1BB costimulatory domain and a 1D9-derived scFv. The resulting P2RX5-directed CAR T cells were non-fratricidal, even when generated from donor T cells homozygous for P2RX5-FL. Furthermore, they showed no reactivity with autologous T cells, Nectin-4-positive cell lines, primary keratinocytes, and cell lines that were either P2RX5-FL haplotype-negative or had undergone P2RX5 gene knockout, consistent with a lack of off-tumor toxicity. Most importantly, the resulting P2RX5-directed CAR T cells displayed potent anticancer activity in both cell cultures and patient-derived xenografts of MM and T-ALL, including heterozygous tumors expressing only one copy of P2RX5-FL.
In summary, we have developed CAR T cells targeting the full-length P2RX5 variant expressed in most patients of African ancestry. These CAR T cells were non-fratricidal and demonstrated high preclinical efficacy against both B- and T-cell malignancies of the P2RX5-FL genotype. This makes them strong candidates for clinical development.
DISCLOSURES
Zhiwei Ang, PhD: No relevant conflicts of interest to declare.