Extracellular domain, hinge, and transmembrane determinants affecting surface CD4 expression of a novel anti-HIV chimeric antigen receptor (CAR) construct

Abstract

Chimeric antigen receptor (CAR)-T cells have demonstrated clinical potential, but current receptors still need improvements to be successful against chronic HIV infection. In this study, we address some requirements of CAR motifs for strong surface expression of a novel anti-HIV CAR by evaluating important elements in the extracellular, hinge, and transmembrane (TM) domains. When combining a truncated CD4 extracellular domain and CD8α hinge/TM, the novel CAR did not express extracellularly but was detectable intracellularly. By shortening the CD8α hinge, CD4-CAR surface expression was partially recovered and addition of the LYC motif at the end of the CD8α TM fully recovered both intracellular and extracellular CAR expression. Mutation of LYC to TTA or TTC showed severe abrogation of CAR expression by flow cytometry and confocal microscopy. Additionally, we determined that CD4-CAR surface expression could be maximized by the removal of FQKAS motif at the junction of the extracellular domain and the hinge region. CD4-CAR surface expression also resulted in cytotoxic CAR T cell killing of HIV Env+ target cells. In this study, we identified elements that are crucial for optimal CAR surface expression, highlighting the need for structural analysis studies to establish fundamental guidelines of CAR designs.

Fig 1. Schematics representing the CAR vector maps for each of the major constructs used in this study.

CD4 extracellular domain is designated as domain 1 and 2 (yellow), and domain 3 and 4 (orange). Black hinge and TM indicate a CD28 origin; red hinge and TM indicate CD8α origin. Intracellular domains are indicated at the top. Vectors were named based on extracellular configuration, length of hinge, transmembrane domain used, and motifs included.

Fig 2. Novel D1D2 CAR is not expressed on the cell surface but the protein is synthesized.

A. Representative flow cytometry plots of CD4 CAR expression on the surface of vector-transduced HEK293 T cells. In contrast to the C.39.28 CAR vector, no surface expression was detected on cells transduced with D.66.α. B. Representative flow cytometry plots of CD4 CAR expression on the surface of HEK293 T cells transfected by calcium-phosphate method. No surface expression was detected on cells transfected with D.66.α. C. Representative flow cytometry plots of CD4 CAR expression on the surface of vector-transduced HEK293 T cells (x axis) and GFP expression from the second gene (y axis). Dual CD4 CAR and GFP expression was observed with positive control C.39.28-GFP but not with D.66.α-GFP. D. Histogram representing the frequency of transfected HEK 293T cells expressing surface CD4 CAR-GFP in 3 experiments from Fig 2C. Statistical analysis was performed as unpaired parametric two-sample t test; a significant difference in intracellular detection was observed between groups (** = p ≤0.01).

Fig 3. D1D2 CAR is detected intracellularly but not on cell surface.

A. Representative flow cytometry plots of CD4 CAR intracellular expression in transfected HEK293 T cells. Reduced but significant intracellular expression of D.66.α was detected compared to C.39.28 vector. B. Histogram representing the frequency of transfected HEK 293T cells expressing surface CD4 CAR (red filled triangle) and intracellular CD4 CAR (blue circle) expression. Significant difference in CAR detection was observed between the vectors (p<0.0001). No significant difference in surface and intracellular-expressing cells was detected for the positive control C.39.28 (p = 0.191). Significant difference between surface and intracellular expression was observed in the D.66.α CAR (p = 0.0066). C&D. Scatter plot for CD4-CAR expression in transfected HEK 293T cells when monocistronic and bicistronic vectors are grouped. Robust detection of C.39.28 was observed on the surface and intracellularly (p = 0.098). Significant, but comparatively reduced detection of intracellular D.66.α expression was observed with no surface expression (p = 0.0002). E. Images of CAR-transfected HEK 293T cells taken by confocal microscopy. Blue is TO-PRO-3 representing nuclear stain; green is CD4-CAR; and red is caveolin representing the plasma membrane. C.39.28 CAR staining shows green expression above the red stain and around the cell, suggesting surface expression. D.66.α CAR staining shows green expression surrounded by red, suggesting intracellular expression only.

Fig 4. Shortening of the CD8α hinge length and LYC motif are required for D1D2 CAR surface expression.

A. Histogram representing the frequency of transfected HEK 293T cells expressing surface CD4 CAR and intracellular CD4 CAR for the various vectors. Positive CAR detection was observed on the surface of cells transfected with a shorter hinged CAR (p = 0.01); however, expression was still reduced compared to control CAR (p = 0.002). Intracellularly, there was no change in CAR detection between long and short hinge (p = 0.227). Enhance frequency of CAR surface detection was observed on cells transfected with CAR that included the LYC motif (p = 0.05); however, expression was significantly reduced if the construct containing a LYC motif also included a longer CD8α hinge (p = 0.02). B. Histogram representing the frequency of HEK 293T transfected with CARs containing the longer CD8α hinge to eliminate any variability/benefit given by a shorter CD8α hinge. When LYC motif was mutated to TTA or TTC, total CAR expression was significantly reduced (p = 0.01 and p = 0.03) and more intracellular CAR expression than surface expression (p = 0.04). C. Images of CAR-transfected HEK 293T cells taken by confocal microscopy. Blue represents TO-PRO-3 nuclear staining; green represents CD4-CAR staining; and red represents caveolin plasma membrane staining. Differences in confocal surface CAR detection were observed, with D*.66.α^LYC expressed on the cell surface but D*.66.α^TTA and D*.66.α^TTC confined intracellularly. Statistical analyses were done using unpaired parametric two-sample t-test.

Fig 5. FQKAS motif hinders but does not inhibit CAR surface expression.

Histogram representing the proportion of transfected HEK 293T cells expressing surface CD4 CAR and intracellular CD4 CAR. Positive CAR expression was observed on the surface of all transfected cells, but the frequency was significantly higher in vectors that did not incorporate FQKAS (p = 0.006). Statistical analysis was done using unpaired parametric two-sample t-test.

Fig 6. Summary graphic of variant CAR surface expression.

Data analysis was based on the relative surface expression of positive control, whole CD4 CAR. Addition of LYC motif and shortening CD8α hinge each partially recovered CAR surface expression, which was enhanced when both modifications were combined. Removal of motif FQKAS improved CAR surface expression to maximal levels seen in the positive control.

Fig 7. Cytolytic activity of CAR T cells.

Primary human T cells were transduced with select vectors and incubated for 24 hours with HIV+ target cells at different E:T ratios. Untransduced T cells showed limited cytotoxic abilities whereas T cell transduced with D.45.α^Lyc showed similar killing capacity as the full length CD4 CAR C.39.28.