Clinical Scale Zinc Finger Nuclease-mediated Gene Editing of PD-1 in Tumor Infiltrating Lymphocytes for the Treatment of Metastatic Melanoma

Abstract

Programmed cell death-1 (PD-1) is expressed on activated T cells and represents an attractive target for gene-editing of tumor targeted T cells prior to adoptive cell transfer (ACT). We used zinc finger nucleases (ZFNs) directed against the gene encoding human PD-1 (PDCD-1) to gene-edit melanoma tumor infiltrating lymphocytes (TIL). We show that our clinical scale TIL production process yielded efficient modification of the PD-1 gene locus, with an average modification frequency of 74.8% (n = 3, range 69.9-84.1%) of the alleles in a bulk TIL population, which resulted in a 76% reduction in PD-1 surface-expression. Forty to 48% of PD-1 gene-edited cells had biallelic PD-1 modification. Importantly, the PD-1 gene-edited TIL product showed improved in vitro effector function and a significantly increased polyfunctional cytokine profile (TNFα, GM-CSF, and IFNγ) compared to unmodified TIL in two of the three donors tested. In addition, all donor cells displayed an effector memory phenotype and expanded approximately 500-2,000-fold in vitro. Thus, further study to determine the efficiency and safety of adoptive cell transfer using PD-1 gene-edited TIL for the treatment of metastatic melanoma is warranted.

Figure 1

Schematic representation of the modified tumor infiltrating lymphocytes (TIL) process and zinc finger nuclease (ZFN) target site at the PD-1 locus. (a) Schematic of the ZFN target site within the PD-1 gene at the 2q37.3 location. Each ZFN comprises five zinc finger domains that recognize a composite 15bp target site (in red). Each ZFP is coupled to an obligate heterodimer mutant of the FokI endonuclease that mediates dimerization dependent DNA cleavage. (b) Work-flow of large clinical scale RNA electroporation with PD-1 targeted ZFN. TIL were induced to rapidly expand (rapid expansion protocol (REP) day 0) and cultured in a G-Rex100 flask at 37 °C and 5% CO₂. On day 7 of the REP, TIL were harvested and washed two times with Hyclone Electroporation Buffer. Cells were then counted and resuspended in electroporation buffer at a concentration of 1 × 10⁸/ml. Cells were mixed with 120 μg/ml of PD-1 ZFN RNA and transferred to the CL-2 processing assembly for electroporation on the Maxcyte GT Flow Transfection System.

Figure 2

Efficiency of PD-1 gene disruption following electroporation with mRNA encoding PD-1-targeted zinc finger nucleases (ZFNs). Cel-1 surveyor nuclease assay performed 3 days following electroporation. ZFN-dependent gene editing occurred in 46.0% of alleles when tumor infiltrating lymphocytes from Donor 1 were electroporated using a smaller scale cuvette and 42.0% when electroporated using the flow electroporation processing assembly. Gene editing occurred in 37.0% of alleles in Donor 2 and 37.0% in Donor 3.

Figure 3

Expression of PD-1 is reduced following zinc finger nuclease-mediated gene editing. (a) PD-1 cell surface expression by FACs analysis is shown for Donor 2. Results are representative of all three donors. (b) Cell surface expression of PD-1 is reduced at baseline and following anti-CD3/CD28 bead stimulation in tumor infiltrating lymphocytes (TIL) following PD-1-targeted gene editing (TIL/PD-1) compared to untransfected TIL (TIL) and GFP-transfected TIL (TIL/GFP). Percent CD3^+/PD-1⁺ cells for each condition is presented as the mean ± SEM for all three donors. (c) Time course of PD-1 mRNA expression in GFP transfected (TIL/GFP) and gene edited TIL (TIL/PD-1 KO) for Donor 2 during anti-CD3/CD28 bead stimulation analyzed by RT-PCR.

Figure 4

The effect of the flow electroporation protocol and PD-1 gene editing on tumor infiltrating lymphocytes (TIL) proliferation during rapid expansion protocol (REP). (a) Both GFP-transfected (TIL/GFP) and gene-edited TIL (TIL/PD-1) sustained a reduction in fold proliferation when rapid expansion was interrupted by the electroporation process (day 7). (b) A second REP revealed no consistent difference between TIL, TIL/GFP, and TIL/PD-1 KO with respect to fold expansion/proliferative capacity.

Figure 5

Gene edited tumor infiltrating lymphocytes (TIL) display a similar phenotype as unmodified cells and GFP-transfected TIL. FACs analysis was performed of each condition for all three donors. The majority of T cells within the bulk TIL had an effector memory-like phenotype (CD45RA^-/CD62L^-) with slight differences in number of cells expressing higher levels of CD45RA that were independent of gene editing.

Figure 6

Enhanced polyfunctional effector cytokine profile of PD-1 knockout tumor infiltrating lymphocytes (TIL) in vitro. Coculture with melanoma cell lines was performed on day 12 of the rapid expansion protocol, 5 days after electroporation, and cytokine (CXCL8/IL-8, GM-CSF, IFNγ, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12 p70, TNFα, and VEGF) levels were measured using the Luminex Performance Human High Sensitivity Cytokine Panel. Donor 1 showed a polyfunctional effector phenotype and a significant increase in TNFα, GM-CSF, and IFNγ secretion following PD-1 gene editing (black bars) as compared to the unmodified TIL (P < 0.0001, hatched bars). Donor 1 PD-1 KO TIL also showed significantly less inhibition following coculture with mel624/PD-L1, a PD-L1-overexpressing cell line. Similar results seen for Donor 2; however, no significant difference in IFNγ secretion was observed when compared to unmodified TIL with the exception of Donor 2 PD-1 KO TIL cocultured with mel526. Shown are representative data from Donors 1 and 2, one-way analysis of variance, P ≤ 0.0001).