Engineered SH3-Derived Sherpabodies Function as a Modular Platform for Targeted T-cell Immunotherapy

Abstract

Engineered T-cell therapies have emerged as a promising approach for cancer treatment, yet their application to solid tumors remains challenging because of the limited specificity and persistence of current antigen recognition strategies. In this study, we introduce sherpabodies, engineered from a human SH3 domain scaffold, as a class of antibody-mimetic proteins capable of precise tumor-associated antigen (TAA) recognition. A phage display library identified sherpabodies against a panel of popular TAAs, which were subsequently incorporated into second-generation chimeric antigen receptor (CAR) constructs that were termed sherpabody-guided CARs (SbCAR). These SbCARs demonstrated potent in vitro specificity and cytotoxicity against solid cancer TAAs, without cross-reactivity to closely related proteins. The modularity, versatility, and small size of sherpabodies enabled generation of multispecific SbCARs, in particular trispecific SbCARs with OR logic that could robustly activate with cells expressing any or combinations of three cognate TAA targets, as well as circuits with IF-THEN logic in combination with synthetic Notch. In vivo, SbCAR T cells elicited a dose-dependent antitumor response in xenograft mouse models, highlighting their potential for therapeutic application. Furthermore, an inducible SbCAR system displayed enhanced persistence and antitumor activity when compared with constitutive CARs. These findings suggest that sherpabodies represent a versatile and promising platform for the next generation of CAR T-cell therapies, particularly for solid tumors. Significance: Sherpabodies represent a biological targeting technology that could help extend the success of CAR T-cell therapy from treating leukemias and lymphomas to the treatment of solid cancers.

Figure 1.. Platform for selection and characterization of sherpabodies targeted against common tumor-associated antigens in solid tumors.

A. Structure of the 60 aa human NPHP1 (nephrocystin) SH3 domain-derived sherpabody targeting scaffold. The five β-strands giving rise to its beta-barrel fold are indicated in purple. The random hexamer sequences used to replace six and three NPHP1 residues in the loop regions before and after, respectively, of the second β-strand of the are shown in red. B. Schematics of the sherpabody (Sb) selection strategy combining phage panning against immobilized tumor associated antigens. Three or four rounds of selection yielded high affinity sherpabody clones that were sequenced and further characterized by ELISA. C. Selection of 20 Sb clones obtained from biopanning procedure that were further evaluated. D. Semiquantitative phage-ELISA of each sherpabody to the panel of immobilized tumor associated antigen. Data was normalized using the interaction of sherpabody-fused E-tag and rabbit polyclonal E-tag antibody. E. Binding affinities for anti-HER2 Sbs as determined by biolayer interferometry. Sensograms show the binding kinetics for human HER2 and immobilized anti-HER2 Sbs. Data are shown as colored traces and the best fit for data to a 1:1 model is shown in dashed lines. The HER2 concentrations used for the binding affinity measurements are indicated.

Figure 2.. In vitro activity of sherpabody-guided CAR constructs in Jurkat and human primary CD8 T cells

A. Schematics of J76-TPR Jurkat T cell-derived reporter line, which expresses a fluorescent protein under the control of T cell activation-induced transcription factors. The image depicts GFP controlled by the NFAT promoter. Representative flow cytometry distributions of GFP reporter for the J-CAR T cells engineered with anti-HER2 SbCARs and cultured with target proteins immobilized on cell culture plastic. B. Percent of activated T cells as determined by GFP expression of J-CAR T cells engineered with sherpabody CARs and cocultured with a panel of target proteins immobilized on cell culture plastic. The color scale has been adjusted to the maximum % of activated cells and the percent of activated T cells was calculated relative to the GFP fluorescence of a control SbCAR (Sb911) that does not target any antigen. Unst = Unstimulated T cells. C. Representative FACS distributions for GFP expression of J-CAR T cells reporter line engineered with SbCARs targeting the selected tumor associated antigens and cocultured with 293T* (HER2 KD, see Figure S3) transfected with a panel of tumor associated antigens. The percent of activated T cells is indicated and was calculated using T cells alone as a reference. D. Representative FACS distributions for GFP expression of J-CAR T cells reporter line engineered with SbCARs targeting the selected tumor associated antigens and cocultured with a panel of cancer cell lines (BT474, HeLa, PC3). The Percent of activated T cells is indicated and was calculated using T cells alone as a reference. E. In vitro killing activity, measured by target cell bioluminescence, for human primary CD8+ T cells engineered with indicated SbCARs and co-cultured with a panel of cancer cell lines. Data are shown as % target survival with bars that indicate the mean and standard error from the mean (n=3) at effector to target ratio 1:1.

Figure 3.. Activation of multispecific SbCARs with OR and IF-THEN logic.

A. Schematics of a trispecific SbcAR with an OR-logic configuration. Representative flow cytometry distributions of GFP reporter for the J-CAR T cells engineered multispecific SbCARs and cultured with target proteins immobolized on cell culture plastic. The histograms are stacked by target protein. B. Percent of activated T cells as determined by GFP expression driven by a NFAT promoter in J-CAR T cells. The composition and specificity for each trispecific sherpabody CAR is indicated and color coded by antigen. The percent of activated T cells was calculated relative to the GFP fluorescence of SbCAR T cells without targets. The color scale has been adjusted to the maximum % of activated cells. C. Schematics of a trispecific SbcAR with an IF-THEN logic configuration. A synthetic notch (SynNotch) guided by an anti-HER2 scFv controls the expression of a Bispecific SbCAR with an OR-logic configuration. The biSbCAR was fused to an mCherry fluorescent protein for detection. Jurkat T cells where cocultured with 293T* (HER2 KD) cells transfected with the indicated combination of tumor associated antigens. Representative flow cytometry distributions of mCherry expression and GFP expression driven by a NFAT promoter in J-CAR T cells. The corresponding CAR is indicated, top panel corresponds to Sb022-Sb025, and bottom panel to control CAR (Sb911). The percent of activated T cells was calculated relative to the mcherry fluorescence of T cells alone. The percent of activated T cells was calculated from cells presenting high levels of mCherry signal.

Figure 4.. In vitro and in vivo characterization of Sb1205 anti-HER2 CAR.

A. Representative FACS distributions showing the EGFP fluorescence driven by a NFAT promoter in J-CAR T cells engineered with Sb1205 anti-HER2 CAR and co-cultured with a panel of cancer cell lines. The percent of activated T cells was calculated relative to the GFP fluorescence of SbCAR T cells alone (Unst.). B. In vitro cell killing curves as a function of target cell antigen density, using human primary CD8+ T cells expressing a constitutive Sb1205 anti-HER2 CAR. Four different effector-to-target ratio are shown. 20,000 Target cells were cocultured with (20000, 15000,10000 and 7000) engineered T cells. The lines are drawn based on inspection. The percentage of specific lysis was determined using flow cytometry by counting the number of target cells after 3 days relative to a co-culture of targets in the presence of untransduced T cells. Data are shown as the mean and standard error from the mean (n=3). Target cells were K562 cells engineered to express different levels of HER2 C. Tumor volumes of breast cancer expressing high HER2 (HCC1569) after treatment with 8 million (4M CD8 + 4M CD4) T cells expressing a Sb1205 anti-HER2 CAR. The solid lines connect the means, and the error bars are the standard error of the mean (n=5). The black dashed line shows the tumor volumes after treatment with 8 million (4M CD8 + 4M CD4) untransduced T cells. (See Figure. S9 for details of control experiment). 3 million target cells were inoculated into the flanks of NSG mice and T cells were transferred 28 days after tumor inoculation D. Representative image of luciferase signal 9 days after T cell injection in tumor bearing mice treated with Sb1205 CAR T cells transduced with a luciferase reporter. Luciferase signal was only detected in the high HER2 tumor, indicating localized expansion (n=5). E. Tumor volumes of breast cancer expressing high HER2 (HCC1569) after treatment with two different doses of T cells expressing a Sb1205 anti-HER2 CAR. The solid lines connect the means and the error bars are the standard error of the mean (n=5). The black dashed line shows the tumor volumes after treatment with 8 million (4M CD8+ and 4M CD4) untransduced T cells. (See Figure S9 for details of control experiment). 3 million target cells were inoculated into the flanks of NSG mice and T cells were transferred 28 days after tumor inoculation. Statistical longitudinal analyses were performed over entire segments of the tumor growth curves using TumGrowth (34). See methods for more details.

Figure 5.. In vitro and in vivo activity of constitutive and inducible Sb1205 anti-HER2 CARs

A. Schematics of a circuit for inducible expression of a SbCAR. An scFv-guided synNotch receptor is used to control the expression of a SbCAR. Differentiation Phenotype of T cells based on CD45RA and CD62L markers. Percentage of CD62L+CD45RA+ T cells for constitutive and inducible Sb1205 CAR, and untransduced T cells in co-cultures with targets expressing the cognate antigen (HER2) (n=4 per group, four independent donor T cells). The bars show the average of measurements, and the bars indicate the standard error of the mean. The lines connect the corresponding measurements for each donor B. Tumor volumes of ovarian cancer (SKOV3) after treatment with T cells expressing a constitutive (dark orange) or inducible (yellow) Sb1205 anti-HER2 CAR. The solid lines connect the means, and the error bars are the standard error of the mean. The gray line shows the tumor volumes after treatment with untransduced T cells. (See Fig. S13 for details of individual animals n=6–9). Three million high HER2 ovarian cancer (SKOV3) were injected subcutaneously in the flanks of NSG mice. 4 million (2M CD8 + 2M CD4) engineered or untransduced human primary T cells were injected i.v. 37 days after tumor injection. C. Tumor volumes of ovarian cancer (SKOV3) after treatment with T cells expressing a constitutive or inducible Sb1205 anti-HER2 CAR, T cells from an additional donor. The solid lines connect the means, and the error bars are the standard error of the mean. The gray line shows the tumor volumes after treatment with untransduced T cells. 3 million high HER2 ovarian cancer (SKOV3) were injected subcutaneously in the flanks of NSG mice. 8 million (4M CD8 + 4M CD4) untransduced or constitutive Sb1205 anti-HER2 CAR human primary T cells or 4 million (2M CD8 + 2M CD4) inducible Sb1205 anti-HER2 CAR were injected i.v. 37 days after tumor injection. The black arrow indicates the time (day 75) when tumors from the constitutive and inducible SbCAR T cells were collected. Statistical longitudinal analyses were performed over entire segments of the tumor growth curves using TumGrowth D. Percentage of CD3+ T cells in tumors collected after 75 days post-implantation (n = 3) from mice treated with constitutive or inducible Sb1205 CAR T cells. E. Expression of exhaustion markers on constitutive CAR and inducible Sb1205 CAR T cells, collected after 75 days post-tumor injection. Pie chart shows percentage of cells that express 0, 1, 2, or 3 exhaustion markers (PD1, LAG3, or TIM3), average of three different tumors. Arc length show the percentage of the indicated marker in each population. F. Percentage of cells expressing the exhaustion markers (PD1, LAG3, or TIM3) on T cells collected from tumors after 75 days post-implantation.

Figure 6.. Sherpabodies can serve as a versatile recognition platform for engineered T cells.

A. Schematics and size comparison of a single chain antibody fragment (ScFv) and a sherpabody. B. Construct design for a sherpabody-based and a ScFv-based CAR. Sherpabodies are compact and can save up to a third of genetic payload in a CAR construct. C. Schematics of a multispecific sherpabody-based trimeric recognition domain and multispecific receptors and circuits for OR- and IF-THEN logic gated SbCARs showcased in this study.