Rationally Designed Transgene-Encoded Cell-Surface Polypeptide Tag for Multiplexed Programming of CAR T-cell Synthetic Outputs

Abstract

Synthetic immunology, as exemplified by chimeric antigen receptor (CAR) T-cell immunotherapy, has transformed the treatment of relapsed/refractory B cell-lineage malignancies. However, there are substantial barriers-including limited tumor homing, lack of retention of function within a suppressive tumor microenvironment, and antigen heterogeneity/escape-to using this technology to effectively treat solid tumors. A multiplexed engineering approach is needed to equip effector T cells with synthetic countermeasures to overcome these barriers. This, in turn, necessitates combinatorial use of lentiviruses because of the limited payload size of current lentiviral vectors. Accordingly, there is a need for cell-surface human molecular constructs that mark multi-vector cotransduced T cells, to enable their purification ex vivo and their tracking in vivo. To this end, we engineered a cell surface-localizing polypeptide tag based on human HER2, designated HER2t, that was truncated in its extracellular and intracellular domains to eliminate ligand binding and signaling, respectively, and retained the membrane-proximal binding epitope of the HER2-specific mAb trastuzumab. We linked HER2t to CAR coexpression in lentivirally transduced T cells and showed that co-transduction with a second lentivirus expressing our previously described EGFRt tag linked to a second CAR efficiently generated bispecific dual-CAR T cells. Using the same approach, we generated T cells expressing a CAR and a second module, a chimeric cytokine receptor. The HER2txEGFRt multiplexing strategy is now being deployed for the manufacture of CD19xCD22 bispecific CAR T-cell products for the treatment of acute lymphoblastic leukemia (NCT03330691).

Conflict of interest statement

Conflict of Interest Statement: M.C.J. has interests in Umoja Biopharma and Juno Therapeutics, a Bristol-Myers Squibb company. He is a seed investor and holds ownership equity in Umoja, serves as a member of the Umoja Joint Steering Committee, and is a Board Observer of the Umoja Board of Directors. M.C.J. also holds patents, some of which are licensed to Umoja Biopharma and Juno Therapeutics. M.C.J. and A.J.J. are inventors on issued and pending patents for the HER2t and EGFRt surface tags, and Seattle Children’s Hospital is an applicant. All other authors report no conflicts of interest.

©2021 American Association for Cancer Research.

Figures

Figure 1.. Design and testing of HER2t construct.

(A) The four extracellular domains of HER2 (left). Molecular model of the extracellular region of HER2t (aa 563–652; middle), HER2t in complex with trastuzumab (Herceptin) Fab (right). Models rendered using UCSF Chimera (PDB 1N8Z) (17). HER2t transmembrane model not present. (B) Schematic of the HER2t amino acid sequence. There is a leader peptide (LP), which is composed of a start codon and the GM-CSF receptor-α chain signal sequence (GMCSFRss) to allow for surface expression, a minimal epitope of HER2 Domain IV (D IV; 89aa, aa563–653 of full-length HER2), and a 23aa transmembrane region (TM; aa653–675). (C) Flow cytometric analysis of HER2t-transduced K562 cells pre- and post-selection using biotinylated trastuzumab and anti-biotin microbeads. Cells were purified to up to 95% HER2t+ relative to untransduced background, and stained with biotinylated trastuzumab, followed by 0.5μg PE-conjugated streptavidin (SA-PE), and analyzed by flow cytometry. (D) Titration of biotinylated trastuzumab against a HER2t-expressing K562 cell line. A total of 1×106 cells were stained with 0 to 2.3 ng of biotinylated trastuzumab, followed by 0.5μg SA-PE, and analyzed by flow cytometry. Gray histograms represent untransduced parental K562 cells (bottom, C and D). Data representative of two independent experiments.

Figure 2.. HER2t partners with trastuzumab to immunomagnetically enrich HER2t expressing cells.

(A) HER2t was cloned in-frame downstream of the second generation 41BB-ζ CD19CAR (23) and T2A to allow for co-expression. Transgenes were under the control of the Elongation Factor 1 promoter (EF1p). CD20CAR-T2A-EGFRt was generated using a similar framework. (B) Marker positivity of CD19CAR-expressing CD8+ T cells pre- and post-selection using biotinylated trastuzumab (HER2t) or cetuximab (EGFRt) and anti-biotin microbeads. Percent positive staining and median fluorescence intensity (MFI) is indicated in each histogram. (C) Western blot analysis using a CD3ζ-specific antibody on cell lysates of HER2t- or EGFRt-purified CD8+ T cells. Cells were untransduced (mock) or transduced with CD19CAR-T2A-HER2t or CD19CAR-T2A-EGFRt. Upper bands = CD19CAR, Lower bands = endogenous CD3ζ. Data in figure from a representative donor of three separate donors. Values under lanes indicate expression ratios. Full western blot is in Supplemental Figure S2A.

Figure 3.. CD19CAR-HER2t and CD19CAR-EGFRt transduced CD8+ T cells maintain effector phenotype and target specificity.

(A) 4hr chromium release assay showing CD19CAR-HER2t and CD19CAR-EGFRt CD8+ T-cell specificity against target cells. CD8+ T cells were co-cultured with target cells at a 50:1, 25:1, 12.5:1 or 6.25:1 ratio. Percentage lysis (mean ± SE) of triplicate wells is depicted. (B) 24hr cytokine release assay. CD8+ T cells were co-cultured at a 2:1 ratio T cell to target for 24hrs and then supernatant was analyzed for the presence of effector cytokines IL2, IFNγ, and TNFα. Assays run in triplicate and data shown as mean ± SE. (C) Cytokine data from 2 independent experiments were normalized (fold cytokine release by CD19CAR-EGFRt = 1) and analyzed by Student t test. Similar results were seen for CD4+ T cells (Supplemental Figures S5). Data in figure from a representative donor of two separate donors.

Figure 4.. HER2t selection stringency can be tuned via the incorporation of a membrane proximal linker.

(A) Schematic of HER2t variants. A short leader peptide (LP) sequence from the IgG4hinge (23), CD28 hinge (40), or a Gly-Ser linker (HER2tG) was inserted between Domain IV (DIV) and the transmembrane region of HER2. (B) Flow cytometric and copy number analyses of H9 T cells transduced with the indicated HER2t variant (MOI = 1) and subsequently immunomagnetically enriched using biotinylated trastuzumab and anti-biotin microbeads. 1×106 enriched cells were stained with biotinylated trastuzumab and PE-conjugated streptavidin (SA-PE), and then analyzed by flow cytometry (relative MFI indicated for each histogram) or pelleted and subjected to ddPCR WPRE analysis using two copy genome reference to albumin. (C) Marker positivity of CD19CAR-HER2tG and CD19CAR-EGFRt CD8+ T cells pre- and post-selection using biotinylated trastuzumab (HER2tG) or cetuximab (EGFRt) and anti-biotin microbeads. Percent positive staining is indicated for each scatter plot. (D) Western blot analysis using a CD3ζ-specific antibody on cell lysates of HER2tG or EGFRt purified CD8+ T cells. Upper bands = CD19CAR; Lower bands = endogenous CD3ζ. The ratio of CAR zeta chain band intensity over internal zeta chain band intensity was set to zero for Mock transduced cells. Fold increases in zeta chain ratios set for CAR transduced populations. Full western blot Supplemental Figure S8A; CD4, full western blot Supplemental Figure S8B. Primary T-cell data in figure from a representative donor of three separate donors.

Figure 5.. HER2tG is a complementary epitope for dual-targeted T cells and an efficient target for in vivo tracking.

(A–C) Bulk CD4+ and CD8+ T cells were transduced with CD19CAR-T2A-HER2tG_epHIV7, CD19CAR-T2A-EGFRt_epHIV7, CD20CAR-T2A-EGFRt_epHIV7, or CD19CAR-T2A-HER2tG_epHIV7 and CD19CAR-T2A-EGFRt_epHIV7. (A) Marker positivity of CD4+ and CD8+ T cells transduced with CD19CART2A-HER2tG_epHIV7 and CD19CAR-T2A -EGFRt_epHIV7 pre-selection and post dual-selection using biotinylated trastuzumab (HER2tG) and cetuximab (EGFRt) and anti-biotin microbeads. Percent positive staining is indicated for each scatter plot. (B) 4hr chromium release assay showing cytolytic response of untransduced CD8+ T cells and the transduced CD8+ T cells against Raji parental and CD19 knockout (CD19KO) Raji created by CRISPR editing (see Materials and Methods), and CD19+, CD20+, CD19+/CD20+ and control (K652) target cells. CD8+ T cells were co-cultured with target cells at a 50:1, 25:1, 12.5:1 or 6.25:1 ratio. Graphs are from a representative donor of three donors, and percentage lysis (mean ± SE) of triplicate wells is depicted. (C) Cytokine release assay of supernatants harvested 24hr post CD8+ T cell:target co-culture at a 2:1 ratio. Supernatant from triplicate wells was analyzed for the presence of effector cytokines IL2, IFNγ, and TNFα. Data representative of two separate assays and presented as mean ± SE. (D) A 1:1 ratio of 1×107 CD4+ and CD8+ T cells expressing various constructs were injected IV into NSG mice previously injected SC with 5×106 NSØ-IL15 cells. Lentiviral constructs used to transduce the CD4+ and CD8+ T cells were the same as in (A). Bone marrow was harvested 14 days post T-cell injection and cell suspensions were analyzed by flow cytometry. Scatter plots are representative of 2×107 end gated Live (Mock), HER2tG+ (CD19CAR-HER2tG), EGFRt+ (CD19CAR-EGFRt or CD20CAR-EGFRt), or dual HER2tG++EGFRt+ (CD19CAR-HER2tG+CD20CAR-EGFRt) cells found in the bone marrow. Gating strategy Supplemental Figure S9. Data representative of two separate experiments (4 mice per group in one experiment, 3 mice per group in second).

Figure 6.. Multiplexed CAR expression afforded by HER2tG and EGFRt selection overcomes antigen escape in vivo.

(A) NSG mice 2 and 8 days post tumor inoculation (0.5e6 Raji GFP:ffluc) were injected with a 1:1 ratio of 1×107 CD4+ and CD8+ T cells transduced with various constructs. CAR T cells were untransduced (Mock), CD19CAR -HER2tG, CD19CAR -EGFRt, or CD20CAR- EGFRtCD19CAR -HER2tG. Serial bioluminescence tumor imaging and Kaplan-Meier analysis of survival were performed for treatment and control mouse cohorts. Grey lines in imaging graphs represent individual mice, 5/group. ****P < .0001 (B) Cohorts of NSG mice were inoculated IV with 0.5e6 cells, 30% of the cells were CD19KO GFP:ffluc Raji and 70% GFP:ffluc Raji on Day 0. The mice were treated on Day 2 and 8 with a 1:1 ratio of 1×107 CD4+ and CD8+ CAR T cells, and serially imaged and analyzed by Kaplan-Meier as in A. ** P < .003 relative to CD20CAR -EGFRt+CD19CAR-HER2tG treated mice. ns = not significant. (C) Flow cytometric analysis and relative representation of CD19– and CD19+ Raji eGFP:ffluc from pre-injection cells or Mock treated, CD19CAR-HER2tG or CD19CAR-EGFRt treated, or CD20CAR-EGFRt+CD19CAR-HER2tG treated mice (as in B). Scatter plots are representative of 2×107 end-gated Live/CD45+/GFP+ tumor cells found in the bone marrow. Gating strategy Supplemental Figure S10.

Figure 7.. HER2tG is an enrichment marker for multiplexed vector platforms.

(A) Growth kinetics of untransduced (Mock), CD19CAR-EGFRt, anfd CD19CAR-EGFRt+CCR(CD122)-HER2tG transduced CD8+ T cells. Transduced cells were purified by EGFRt on Day 11 of culture then separated into triplicate wells in the presence or absence of cytokine. Viability and proliferation were subsequently measured using the NucleoCounter® NC-3000. (B) Flow analysis of HER2tG marker expression Day 15 and Day 30 post stimulation with CD3/CD28. A subset of CD8+ T cells from A (–IL-2 & IL-15) was subjected to flow analysis using biotinylated trastuzumab, PE-conjugated streptavidin (SA-PE), and APC-conjugated cetuximab. Percent positive HER2tG staining is indicated for each histogram (middle, right). Left, representative CAR-positivity as evidenced by EGFRt staining. Data in figure from a representative donor of two separate donors.