Arming Immune Cell Therapeutics with Polymeric Prodrugs

Abstract

Engineered immune cells are an exciting therapeutic modality, which survey and attack tumors. Backpacking strategies exploit cell targeting capabilities for delivery of drugs to combat tumors and their immune-suppressive environments. Here, a new platform for arming cell therapeutics through dual receptor and polymeric prodrug engineering is developed. Macrophage and T cell therapeutics are engineered to express a bioorthogonal single chain variable fragment receptor. The receptor binds a fluorescein ligand that directs cell loading with ligand-tagged polymeric prodrugs, termed "drugamers." The fluorescein ligand facilitates stable binding of drugamer to engineered macrophages over 10 days with 80% surface retention. Drugamers also incorporate prodrug monomers of the phosphoinositide-3-kinase inhibitor, PI-103. The extended release of PI-103 from the drugamer sustains antiproliferative activity against a glioblastoma cell line compared to the parent drug. The versatility and modularity of this cell arming system is demonstrated by loading T cells with a second fluorescein-drugamer. This drugamer incorporates a small molecule estrogen analog, CMP8, which stabilizes a degron-tagged transgene to provide temporal regulation of protein activity in engineered T cells. These results demonstrate that this bioorthogonal receptor and drugamer system can be used to arm multiple immune cell classes with both antitumor and transgene-activating small molecule prodrugs.

Keywords: cell backpacking; cell-mediated targeting; macrophages; polymeric prodrugs; solid tumors.

Figure 1.

Schematic mechanism of polymeric prodrug ‘drugamer’ loading on engineered cell therapeutics. Immune cells are engineered with a biorthogonal and humanized scFv receptor that binds to polymeric prodrugs via a high affinity receptor-ligand binding reaction. The polymeric prodrugs sustain the release of small molecule drug cargo via tunable linker design and selection. The versatility of this platform has been demonstrated by arming genetically engineered macrophage cells with polymeric prodrugs of a PI3K kinase inhibitor PI-103 (upper right), and by arming engineered T cells with a different drugamer that releases CMP8 to switch on and off the model degron-tagged eBlue Fluorescent Protein 2 activity (lower right).

Figure 2.

Kinase inhibitor drugs including PI-103, BKM-120, GSK-2636771, and Dasatinib were screened for toxicity against human monocyte-derived macrophages and normalized to DMSO vehicle. Viability was measured as a function of relative ATP production. PI-103 (blue circle); BKM-120 (green triangle); GSK-2636771 (red square); Dasatinib (pink diamond). Data are presented as mean ± SD, baseline corrected (n = 4).

Figure 3.

Synthesis and characterization of PI-103 drugamer. (A) Drugamer synthesis by Reversible Addition Fragmentation-chain Transfer (RAFT) polymerization with 4-cyano-4-(phenylcarbonothioylthio)-pentanoic acid (CTP) and 4,4′-azobis(4-cyanovaleric acid) (ABCVA) and by incorporation of methacrylate monomers: polyethylene glycol, molecular weight 950 (PEGMA950); PI-103 kinase inhibitor (blue); and fluorescein ligand for assembly and binding to scFv receptor (green); labeled with rhodamine (pink). (B) PI-103 release kinetics from drugamer over 22 days incubated at 37 °C in 50% human serum. Experiment performed in triplicate and data expressed as mean with standard deviation bars. (C) PI-103 inhibition of phospho-Akt in U87 cells was characterized following 24h incubation with parent PI-103 and the PI-103 drugamer and compared to control untreated cells.

Figure 4.

Lentiviral transduction of primary human macrophages provides high and bioorthogonal expression of AntiFl-ζ. (A) The AntiFl-ζ expression construct is composed of an anti-flourescein scFv (FITC-E2) linked by a glycine repeat to an EGFRt transmembrane domain and surface tag. The control construct consists of a truncated CD19 tag without FITC-E2. Constructs were cloned into an epHIV7.2 lentiviral backbone and co-transfected into 293T cells with packaging plasmids for synthesis of lentiviral particles (LPs). Transduction of patient-derived macrophages yields the AntiFl-ζ GEMs with the bioorthogonal surface receptors for drugamer assembly (B) AntiFl-ζ expression by GEMs occurs with high transduction efficiency around 250 LPs/cell (99% FITC-E2+) and increases with increasing LPs/cell. Shown are representative plots from one donor (n = 3). (C) Comparison of untransduced, wild type macrophages, CD19t GEMs and AntiFl-ζ GEMs expression of characteristic macrophage phenotype markers (CD40, CD80, CD86, HLA-DR/DP/DQ, CD11b, PD-L1, CD163, CD206, and CD209) 7 days after lentiviral transduction shows minimal changes to cell phenotype following lentiviral transduction. Shown are representative results from one donor (4 patient donor-derived macrophages were studied in total).

Figure 5.

Loading and characterization of genetically engineered macrophages (GEMs) with rhodamine-labeled drugamer (A) Cell loading was measured with anti-fluorescein scFv (AntiFl-ζ) and control truncated CD19 (CD19t) GEMs (no scFv) as a function of drugamer concentration to determine the binding curve (AntiFl-ζ GEMs, green circle; CD19t GEMs, blue square; n = 3, representative plot corrected against baseline fluorescent levels (0.527 × 10³ MFI)) (B) PI-103 drugamer remains bound to GEMs for up to 10 days in cell culture as evidenced by retention of rhodamine fluorescence. Data shown are representative results from one donor (n = 3; 2^nd and 3^rd replicates conducted with control drugamer). (C) The level of PI-103 drugamer internalization by the AntiFl-ζ GEMs was measured after acid-salt washout of surface-bound polymer and determination of rhodamine (Rh) fluorescence loss. Results determined that approximately 80% of drugamer remained bound on the AntiFl-ζ GEMs over ten days. Data shown are representative results from one donor (n = 3).

Figure 6.

The longitudinal anti-U87 glioblastoma tumor cell activity of the PI-103 drugamer was compared to parent PI-103 drug, to control drugamer that does not contain PI-103, and to nontreated and nocodazole (NOC; a potent mitotic inhibitor drug) controls. The PI-103 drugamer was incubated at 5 μM polymer (corresponding to 30 μM PI-103 at 100% release), and significantly inhibited U87 ATP production for the five days characterized compared to parent PI-103. The control drugamer had no significant inhibitory effect on cells. Average bioluminescence per well was measured in counts per second (CPS) and scaled with relative ATP production for each U87 treatment group. Data are presented as mean ± SD, baseline corrected (n = 4). Welch’s t-tests were used to determine statistical significance between groups at specific time point (*p < 0.05, **p < 0.01, and ***p < 0.001).

Figure 7.

Arming engineered T cells with CMP8 drugamer prevents degron-tagged degradation of intracellular eBFP2 and extends the protein fluorescence activity. (A) The T cell line, H9, was first used to validate the versatility of AntiFl expression and function in different immune cell classes. AntiFl H9s were loaded with CMP8 drugamer and drugamer fluorescence was compared against drugamer-loaded wild type (WT) and AntiFl H9 controls with no drugamer loading (n =1, 28,311 ± 12,390 cells total analyzed per group). Figure S8 shows loading of Jurkat T cells similarly. (B) Jurkat T cells were co-transduced with AntiFl and a transgene encoding for a degron-tagged fluorescent protein, eBFP2 (AntiFl.BFP Jurkats). Arming of AntiFl.BFP Jurkats with CMP8 drugamer resulted BFP fluorescence detection induced by the gradual release of CMP8 from the drugamer and compared against a CMP8 drugamer loaded WT Jurkat control. eBFP2 fluorescence detection from CMP8 drugamer-loaded AntiFl.BFP Jurkats increased steadily over 24 hours (n =1, 21,559 ± 10,833 cells total analyzed per group, %eBFP2 positivity normalized against 0h WT Jurkat).