Using quantitative single molecule localization microscopy to optimize multivalent HER2-targeting ligands

Abstract

Introduction: The progression-free survival of patients with HER2-positive metastatic breast cancer is significantly extended by a combination of two monoclonal antibodies, trastuzumab and pertuzumab, which target independent epitopes of the extracellular domain of HER2. The improved efficacy of the combination over individual antibody therapies targeting HER2 is still being investigated, and several molecular mechanisms may be in play: the combination downregulates HER2, improves antibody-dependent cell mediated cytotoxicity, and/or affects the organization of surface-expressed antigens, which may attenuate downstream signaling.

Methods: By combining protein engineering and quantitative single molecule localization microscopy (qSMLM), here we both assessed and optimized clustering of HER2 in cultured breast cancer cells.

Results: We detected marked changes to the cellular membrane organization of HER2 when cells were treated with therapeutic antibodies. When we compared untreated samples to four treatment scenarios, we observed the following HER2 membrane features: (1) the monovalent Fab domain of trastuzumab did not significantly affect HER2 clustering; (2) individual therapy with either trastuzumab or (3) pertuzumab produced significantly higher levels of HER2 clustering; (4) a combination of trastuzumab plus pertuzumab produced the highest level of HER2 clustering. To further enhance this last effect, we created multivalent ligands using meditope technology. Treatment with a tetravalent meditope ligand combined with meditope-enabled trastuzumab resulted in pronounced HER2 clustering. Moreover, compared to pertuzumab plus trastuzumab, at early time points this meditope-based combination was more effective at inhibiting epidermal growth factor (EGF) dependent activation of several downstream protein kinases.

Discussion: Collectively, mAbs and multivalent ligands can efficiently alter the organization and activation of the HER2 receptors. We expect this approach could be used in the future to develop new therapeutics.

Keywords: HER2; meditope; pertuzumab; single molecule localization microscopy; trastuzumab; valency.

FIGURE 1

Effect of mAbs on HER2 nanoscale organization. (A) SMLM images of HER2-paGFP within a region from MDA-MB-468 cells in steady state and upon mAb/Fab treatment. An example for steady state HER2 is shown on the left, and examples for the organization of HER2 upon 10-min treatment with 20 nM Fab trastuzumab, 10 nM trastuzumab, 10 nM pertuzumab, and co-treatment with 10 nM trastuzumab and 10 nM pertuzumab are shown from left to right. All treatments were performed at 37°C. Scale bars: 200 nm. Standard PALM image analysis was employed (68) and localizations were grouped using a maximum blinking time of 5 s for paGFP and group radius of three times the maximum localization precision. (B) Distribution of detected HER2-paGFP molecules per cluster in steady state and upon mAb treatment. (C) Distribution of cluster radius of HER2-paGFP in steady state and upon mAb treatment. (D) Fraction of clustered HER2-paGFP molecules with SEM; *denotes p value ≤ 0.05. Quantitative analysis (B–D) was based on the following cell and region of interest (ROI) statistics: steady state (12 cells, 26 ROI), Fab trastuzumab (12 cells, 26 ROI), trastuzumab (13 cells, 30 ROI), pertuzumab (13 cells, 36 ROI), and co-treatment with trastuzumab and pertuzumab (12 cells, 29 ROI). ROIs for analysis were 10-18 μm². While Fab trastuzumab had minimal effect on the HER2 distribution, mAbs and mAb combination induced significant clustering.

FIGURE 2

Effect of multivalent meditope-based ligands on HER2 nanoscale organization. (A) SMLM images of HER2-paGFP from a region on MDA-MB-468 cells upon treatment with trastuzumab memAb (TmemAb) in combination with Fc or Fc-multivalent meditiope constructs. Image pairs from left to right show HER2 upon treatment with 10 nM TmemAb for 10 min and subsequent incubation for either: 3 or 10 min with Fc; Fc-divalent meditope (M2Fc); or Fc-tetravalent meditope (M4Fc). All treatments were performed at 37°C and cells were briefly washed with warm media before treatment with Fc or Fc-meditope constructs. Scale bars: 200 nm. Standard PALM image analysis was employed (68) and localizations were grouped using a maximum blinking time of 5 s for paGFP and a group radius of three times the maximum localization precision. (B) Distribution of detected HER2-paGFP molecules per cluster upon treatment with TmemAb in combination with Fc or Fc-multivalent meditiope treatment. (C) Distribution of cluster radius of HER2-paGFP upon treatment with TmemAb in combination with Fc or Fc-multivalent meditiope treatment. (D) Fraction of clustered HER2-paGFP molecules with SEM; *denotes p value ≤ 0.05. Quantitative analysis (B–D) was based on the following cell and region of interest (ROI) statistics: TmemAb and Fc treatment for 3 min (12 cells, 27 ROI), TmemAb and Fc treatment for 10 min (12 cells, 26 ROI), TmemAb and M2Fc treatment for 3 min (12 cells, 28 ROI), TmemAb and M2Fc treatment for 10 min (14 cells, 34 ROI), TmemAb and M4Fc treatment for 3 min (12 cells, 30 regions), TmemAb and M4Fc treatment for 10 min (12 cells, 28 ROI). ROIs for analysis were 10–18 μm². Treatment with M4Fc for 3 min had a significant clustering effect on the HER2 distribution.

FIGURE 3

Trastuzumab memAb and multivalent meditopes induce HER2 reorganization. (A) SMLM images of AF647 labeled trastuzumab memAb (TmemAb) targeting HER2 on BT-474 cells. Live cells were incubated with 10 nM TmemAb-AF647, alone or premixed with 10 nM multivalent meditope (i.e., M2Fc and its variants or M4Fc), for 10 min at 37°C and subsequently fixed. Alternatively, cells were stained with 10 nM TmemAb-AF647 postfixation (PF). Scale bars: 200 nm. (B) Average molecular density of HER2 detected with TmemAb; the p values are provided in Supplementary Table 1. (C) Fraction of TmemAb-bound HER2 identified as an isolated receptor (blue), within a cluster of two receptors (red), or as part of a cluster with more than two receptors (gray). (D) Percentage of clustered proteins within a given ROI; the p values are provided in Supplementary Table 2. Quantitative analysis (B–D) used the following cell and region of interest (ROI; 18 μm²) statistics: TmemAb PF (14 cells, 29 ROI), TmemAb (14 cells, 34 ROI), TmemAb + M2Fc10 (16 cells, 40 ROI), TmemAb + M2Fc20 (17 cells, 39 ROI), TmemAb + M2Fc30 (15 cells, 35 ROI), TmemAb + M2Fc (23 cells, 43 ROI), and TmemAb + M4Fc (24 cells, 57 ROI).

FIGURE 4

Effects of multivalent meditopes on cell signaling pathways upon EGF stimulation of HER2-paGFP transfected MDA-MB-468 cells. Cells were treated with 10 nM trastuzumab memAb for 10 min, followed by either 10 nM Fc, M2Fc, M4Fc or Fc for indicated times. Cells were incubated for additional 30 min without or with addition of 10 ng/ml EGF. Alternatively, cells were co-treated with 10 nM trastuzumab and 10 nM pertuzumab for indicated periods of time. After brief wash with warm media, cells were incubated for additional 30 min without or with addition of 10 ng/ml EGF. All treatments were performed at 37°C; cells were briefly washed with warm media before and after treatment with Fc or Fc-meditope constructs. (A) Phosphorylation of EGFR(Y1068), HER2(Y877), and downstream signaling targets Akt(S473) and Erk1/2(p42/44 T202/Y204) was determined using Western blot analysis for cells treated as described above. (B) Relative phosphorylation of Akt, HER2, and EGFR for cells treated as described above was calculated as the ratio of pAkt over total Akt, pHER2 over total HER2 or pEGFR over total EGFR and normalized to the maximum response by Fc at the indicated time points. Results are representative of 5 independent experiments. Colors show Fc treatment in gray, M2Fc treatment in blue, M4Fc in red, and combined parental trastuzumab and pertuzumab in purple. M4Fc has the highest inhibition of EGF mediated phosphorylation.

FIGURE 5

Effects of M4Fc on ADCC in SK-BR-3 cells. ADCC activity was measured in SK-BR-3 cells treated with increasing doses of trastuzumab or trastuzumab memAb in the presence or absence of 10 nM M4Fc for 6 h. Data was compiled from three experiments and analyzed using a 2-way ANOVA (*p < 0.05; **p < 0.01). Data is normalized to the untreated (0 nM antibody) cell signal. Treatment with M4Fc in the presence of trastuzumab memAb shows higher levels of ADCC activity.