Optimizing Conjugation Chemistry, Antibody Conjugation Site and Surface Density in Antibody-Nanogel Conjugates (ANCs) for Cell-Specific Drug Delivery

Abstract

Targeted delivery of therapeutics using antibody nanogel conjugates (ANCs) with a high drug-to-antibody ratio has the potential to overcome some of the inherent limitations of antibody-drug conjugates (ADCs). ANC platforms with simple preparation methods and precise tunability to evaluate structure-activity relationships will greatly contribute to translating this promise to a clinical reality. In this work, using trastuzumab as a model antibody, we demonstrate a block copolymer-based ANC platform that allows highly efficient antibody conjugation and formulation. In addition to showcasing the advantages of using an inverse electron demand Diels-Alder (iEDDA) based antibody conjugation, we evaluate the influence of antibody surface density and conjugation site on the nanogels upon the targeting capability of ANCs. We show that compared to traditional strain-promoted alkyne-azide cycloadditions, preparation of ANCs using iEDDA provides significant higher efficiency which results in shortened reaction time, simplified purification process as well as enhanced targeting towards cancer cells. We also find that through a site-specific disulfide-rebridging method in antibodies offers similar targeting abilities as the more indiscriminate lysine-based conjugation method. The more efficient bioconjugation using iEDDA allow us to optimize the avidity through fine-tuning the surface density of antibodies on the nanogel. Finally, with trastuzumab-mertansine (DM1) antibody-drug combination, our ANC demonstrates superior activities in vitro compared to the corresponding ADC, further highlighting the potential of ANCs in future clinical translation.

Graphical Abstract

INTRODUCTION

Directing therapeutics to the disease site using antibodies and antibody derivatives has been considered a powerful approach to improve the therapeutic efficacy and attenuate off-target side effects.^1–3 The most successful example of these targeted therapies is antibody-drug conjugates (ADCs) with fourteen drugs approved to date by the FDA for clinical use on various types of cancers.^4–6 Even though ADCs provide a wider therapeutic window compared to traditional chemotherapies, current efficacies are far from optimal and further improvements are limited by the drug-to-antibody ratio (DAR)of ADCs that only allows a maximum of ~8 drug molecules per antibody.^7–9 Recently, researchers have increased the maximum DAR to ~20, where hydrophilic polymers have been utilized to mitigate the solubility issues associated with high DAR ADCs.^10–12 Compared to ADCs and these antibody polymer conjugates (APCs), antibody nanogel conjugates (ANCs) offer an excellent option to conveniently improve the DAR by 10²-10⁵ times.^13,14 In addition to significantly increasing drug exposure to the tumor site due to the high DAR, ANCs prevent premature release of the payload because of the enhanced drug stability within nanogel formulations.^15–19 However, for ANCs to be a clinically realistic platform, antibody conjugation efficiency and batch-to-batch consistency must be substantially improved.^20,21 Therefore, developing and optimizing of ANCs formulations with a simple fabrication process, high conjugation efficiency, cellular targeting, and good control over DAR are essential to fully harness the advantages of ANCs.

Small molecule-based targeting ligands such as folic acid or peptides can be efficiently conjugated to polymer backbones or nanoparticles for precise control of surface ligand density and high batch-to-batch consistency.^22,23 Due to the limited stability of proteins and antibodies in organic solvents, ANCs are typically constructed through post-formulation conjugation chemistries such as thiol-maleimide chemistry, 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide (EDC) coupling, and strain-promoted alkyne-azide cycloadditions (SPAACs).^24,25 SPAAC between dibenzocyclooctyne (DBCO) and azide (N₃) is preferably used in most of the ANC constructs due to its robust conjugate stability and mild reaction conditions.^{14,24,26–29} However, considering the low k₂ value of SPAAC for small molecules is ~0.1 M^-1 s^-1, conjugation kinetics between antibodies and nanogels with significantly higher molecular weights would be even slower.³⁰ Indeed, ANC preparation using SPAACs typically requires long conjugation times (>24 h at 4 °C) and the associated inefficiency demands an additional purification step to remove unconjugated antibodies.^31,32 which further increase difficulties in quality control, optimizations, and characterizations.

Minimizing heterogeneity for optimal target engagement is important in ADC development.^33–35 Non-specific modification of antibodies to generate ADCs limits the understanding of the effect of DAR and drug location upon critical properties such as efficacy, safety, pharmacokinetics, and immunogenicity.^36,37 Multiple publications also indicate that antigen binding affinity by randomly-modified ADCs decrease compared to the corresponding unmodified antibody.^38–40 Considering that these parameters are even much less understood in the emerging ANC platform, it is important to understand the significance of site-specific conjugation of antibodies on the surface of nanogels which could potentially create different antibody orientation on the nanogels surfaces. Additionally, note that nanogels offer the opportunity to introduce multiple copies of antibodies on their surface, which should increase the avidity of ANCs to cells due to multivalent interactions.^41–43 However, the increased avidity could also result in decrease in cellular specificity.^44–46 Therefore, strategies that allow precise control of antibody density on nanogels are critical in the development of ANCs.

To assess the impact of the methods of conjugation and the structure of the ANC, we focused on three features. First, we envisaged an approach to utilize the more efficient inverse electron-demand Diels–Alder reaction (iEDDA) between trans-cyclooctene (TCO) and tetrazine (Tz), to improve the conjugation efficiency between antibody and nanogels (Figure 1). ANCs generated from the iEDDA are directly compared with the corresponding SPAAC-based ANCs. Second, to evaluate the impact of the conjugation site and density of antibodies on the surface of the nanogels upon the targeting ability of ANCs, we utilized a previously reported dibromo-pyridazinediones derivatives^47,48 to site-specifically modify the disulfide bridges of antibodies. Third, we compare the ANC with the corresponding ADC in vitro to evaluate the advantages of the ANC platform.

Figure 1.

Schematic illustration of ANC platforms used in this work. Top panel: overall formulation and conjugation processes. Middle panel: polymer synthesis scheme. Bottom panel: chemistries and conjugation sites used for antibody conjugation.

RESULT AND DISCUSSION

Synthesis of azide- and tetrazine-containing block copolymers for nanogel formulation.

To compare the conjugation efficiency of the SPAAC and iEDDA reactions, we targeted block copolymers containing azido or tetrazine moieties respectively at their hydrophilic terminus. Both of these polymers were achieved from a common amine-terminated precursor polymer P0. This polymer was synthesized as previously described by reversible addition−fragmentation chain transfer polymerization (RAFT) of the hydrophobic monomer pyridyl disulfide methacrylate (PDS-MA) using a Boc-amine terminated PEG5k as the macroinitiator, followed by Boc-deprotection.^49,50 The terminal N₃ was then installed on the polymer using the amine handle to generate the azide-terminated polymer P1, as shown in Figure 1. Polymer P2, bearing the tetrazine moiety was synthesized using a similar approach (see supporting information (SI) for characterization details). Polymer P3, without a reactive terminal functionality, was also synthesized in order to ultimately adjust the density of the conjugation handles on the nanogels through a co-assembly strategy. We have previously shown that a 50:50 ratio of P1 and P3 offers self-assembly characteristics for conjugating targeting ligands or antibodies.²² With this ratio, nanogel formulation was achieved in three simple steps: nanoprecipitation assisted by sonication to form the self-assembled structures, disulfide crosslinking of the assembly to form structurally stable nanogels, and removal of crosslinking by-products and unencapsulated drug or dye molecules to purify the nanogels. Nanogels, loaded with 10% DiI dye by mass, were found to be ~60 nm as assessed by dynamic light scattering (DLS). Nanogels, encapsulating 20% DM1, showed a size of ~80 nm.

Comparison of SPAAC and iEDDA in ANC preparation.

With nanogels in hand, to evaluate their conjugation efficiency with antibodies, we used surface lysines as the handles to prepare and characterize TCO- or DBCO-functionalized antibodies. Trastuzumab, a HER2 targeting monoclonal humanized antibody, was used to test this approach. To ensure similar extents of modification for both functionalities, 7.5 equivalents of DBCO-PEG4-NHS or TCO-PEG4-NHS linker were utilized in the reaction. The resultant functionalized antibodies were purified, deglycosylated, and characterized by LC-MS. The average linker to antibody ratio (LAR) was found to be ~4 (Figure 2a). Prior to the experiment, we wanted to project the estimated reaction progress of iEDDA compared to SPAAC. The obtained LAR in Figure 2a was translated to the concentration of DBCO or TCO in the conjugation mixture. Similarly, the concentration of N₃ or Tz was calculated according to the polymer concentration formulated into the nanogels. We simulated the second-order reaction kinetics based on the published k₂ value of DBCO-azide and TCO-Tz.⁵¹ The results indicate that the reaction of TCO-Tz should be completed within the first hour, while the DBCO-N₃ reaction only reached ~60% completion after 24 hours (Figure 2b). To assess the experimental conjugation reaction efficiency, we evaluated the reaction progression by gel electrophoresis at different time points after conjugation. We have previously used this method to evaluate protein encapsulation efficiency.^52,53 This evaluation was performed on ANC with different degrees of conjugation based on weight ratio of antibody-to-polymeric nanogel weight (See detailed calculation in the Materials and Methods section). As shown in Figures 2c and 2d, nearly 90% of the antibody was consumed in the case of NG^Tz and Ab^TCO. In contrast, it took ~72 hours to make the same reaction progress for NG^N3 and Ab^DBCO. Details on the integration data are shown in Figure S10–16. Also, no significant changes in size were observed after antibody conjugation using either of the conjugation chemistries (Figure S4–9). Overall, these studies show that the TCO-Tz based conjugation between nanogels and antibody is significantly faster and more efficient, leaving only a negligible amount of free antibodies remaining in the reaction mixture.

Figure 2.

(a) Antibody linker conjugation and characterization by LC-MS Trastuzumab was modified with equimolar amount of TCO- or DBCO-PEG-NHS, resulting in antibody with similar number of linkers modified on its surface; (b) Simulation of second order reaction kinetics with varied k₂ value predicting reaction progress of DBCO-N₃ versus TCO-Tz conjugation chemistry. The TCO-Tz reaction approached 100% completion in <1 hour, while DBCO-N₃ took much longer time; (c), (d) Conjugation efficiency between antibody and nanogels at various weight ratios using DBCO-N₃ or TCO-Tz click chemistry, calculated based on the band intensity of unmodified antibody in SDS-PAGE gel. The conjugation using the TCO-Tz reaction reached saturation within the first two hours. DBCO-N₃ reaction took more than 72 hours to achieve similar level of conjugation.

ANC preparation protocol optimization.

Removing unreacted, free antibodies from the conjugation mixture is necessary because the free antibodies can compete with the ANCs in binding to the cells, which could reduce the targeting capability of ANCs. Based on the high efficiency of the TCO-Tz conjugation reaction, we hypothesized that this additional antibody removal step is not necessary, and the cell targeting capability of ANCs will not be influenced (Figure 3a). To test this hypothesis, we prepared DiI-encapsulated ANCs using both DBCO-N₃ and TCO-Tz chemistries, followed by removal of unconjugated antibodies using size exclusion chromatography (SEC). The cell-targeting capabilities of both purified and unpurified groups were evaluated on BT474 (HER2+) breast cancer cells and MCF10A (HER2-)cells using flow cytometry. To ensure the same concentration of ANCs before and after purification, we synthesized a Cy7 labeled block copolymer P4. This polymer was co-assembled with P3 and P1 or P2 to obtain Cy7-labeled NG^N3 and NG^Tz (50% P1/P2, 40% P3, and 10% P4, calculated by wt.), which exhibit the same size as their original counterparts without P4. The Cy7-labeled NG^N3 and NG^Tz nanogels were subjected to the DBCO-N₃ and TCO-Tz antibody conjugation to obtain Cy7-ANC^DBCO and Cy7-ANC^TCO respectively and the reaction mixtures were then subjected to SEC for antibody removal. A peak corresponding to unconjugated trastuzumab was seen in the case of Cy7 ANC^DBCO, while no discernible peak was seen for Cy7 ANC^TCO (Figure S21).

Figure 3.

(a) Schematic illustration of ANC preparation protocol using two types of click chemistry. ANC^DBCO needed longer time to reach complete conjugation and required SEC purification to eliminate unconjugated antibody; (b) Cellular targeting of ANCs with and without purification evaluated by flow cytometry in BT474 (HER2+) and MCF10A (HER2-) cells. SEC purification was performed to eliminate competing unconjugated antibody. Due to poor conjugation efficiency, after antibody removal, ANC^DBCO showed improved targeting capability. In contrast, ANC^TCO had better conjugation efficiency, thus SEC purification was not required. The ANC^TCO also demonstrated superior targeting capability.

With this observation, we posited that the difference in the cellular targeting ability of the unpurified and purified ANC would be significant with Cy7-ANC^DBCO because of the competitive inhibition by the free antibodies, while the difference would be negligible with Cy7 ANC^TCO. With the HER2+ BT474 cells, the targetability of Cy7 ANC^DBCO (1:2 Ab:NG weight ratio) improved by 4-fold as shown by the increased median fluorescence intensity (MFI), compared to the unconjugated nanogel (Figure 3b). An additional 3-fold increase in targeting is observed with Cy7 ANC^DBCO after SEC purification. On the other hand, a 20-fold targeting enhancement is observed with Cy7 ANC^TCO (1:2 Ab:NG weight ratio) relative to the unmodified nanogel. As expected, no further improvement in targeting is observed with SEC purification. Similarly, if the ANCs and the observed differences are due to receptor targeting and competitive binding, the extent of cellular uptake and the differences in cellular uptake should be minimal in the HER2-MCF10A cell line. The observed results are indeed consistent with this expectation (Figure 3b).

The results here also highlight the necessity for additional SEC purification for the DBCO-N₃ conjugation reaction, while TCO-tetrazine conjugation obviates this rather laborious step in addition to the added advantage of shorter reaction time (Figure 3a). Another interesting finding is the inherently higher targeting of Cy7-ANC^TCO in comparison to Cy7-ANC^DBCO (Figure 3b). We speculate that this is due to the higher density of antibodies on the Cy7-ANC^TCO arising from the higher efficiency of the iEDDA conjugation reaction.

Evaluation of ANCs with varied antibody density and conjugation site

After optimizing the preparation protocols, we were interested in assessing the impact of the density and antibody conjugation site on the surface of the nanogel upon the targeting ability of ANCs (Figure 4a). Utilizing lysines as the modification handle on the antibodies affords a random distribution of conjugation sites because of the rather ubiquitous presence of lysines. On the other hand, disulfide rebridging methods using dibromo-pyridazinediones derivatives offer more site-defined conjugation sites for antibodies.⁵⁴ To ensure a similar conjugation efficiency between antibody and nanogels, we designed and synthesized a TCO containing dibromo-pyridazinediones derivatives (Figure S22–27) for comparison with the corresponding lysine-based TCO-modified antibody. Successful disulfide-rebridging based modification is supported by gel electrophoresis (Figure 4b and Figure S28) The resultant TCO-functionalized antibody was then purified by SEC, which was deglycosylated and characterized by LC-MS to determine the LAR. The LAR of the rebridged antibodies is around 2.8 (Figure 4c).

Figure 4.

(a) Schematic illustration of the preparation of ANCs using antibodies modified using lysine chemistry or disulfide rebridging chemistry. The conjugation site was speculated to possibly alter the antibody orientation, hence offering binding site availability; (b) Gel electrophoresis analysis of antibody rebridging process using dibromo-pyridazinedione TCO linker; (c) Preparation and characterization of Ab-Disulfide-TCO by LC-MS.

We then prepared three different ANC scaffolds: ANC^N3-DBCO-Lys, ANC^Tz-TCO-Lys, and ANC^Tz-TCO-SS. A weight ratio of 1:2, 1:5, and 1:10 of antibody to polymer was used to prepare 10% DiI encapsulated nanogels with varied extents of surface antibody decoration. The cell-specific targeting flow cytometry results are shown in Figures 5a and 5b where a ligand-density-dependent cell-specific targeting was observed for all three ANCs in BT474 cells and negligible uptake was observed in MCF10A cells. In addition, site-specific ANC^Tz-TCO-SS showed the best cell targeting, and ANC^Tz-TCO-Lys performed significantly better than ANC^N3-DBCO-Lys at all three-antibody densities. These results highlight the importance of both conjugation efficiency and antibody conjugation site on the targeting capability of ANCs. Figures 5c and 5d show the cellular uptake of ANCs evaluated by confocal fluorescence microscopy where ANC^Tz-TCO-Lys and ANC^Tz-TCO-SS show a significant uptake in BT474 cells at both 1:10 and 1:2 antibody to polymer ratio, whereas ANC^N3-DBCO-Lys only show reasonable cellular uptake at the 1:2 antibody. The negligible cellular uptake in MCF10A cells indicates the selectivity of these ANCs towards HER2 receptors (Figure S33).

Figure 5.

Flow cytometry based comparison of cell targeting capability of ANCs prepared with three different conjugation chemistries in (a) BT474 (HER2+) and (b) MCF10A (HER2-) cells; (c), (d) confocal images of ANCs prepared with three different conjugation chemistries in BT474 cells at 1:2 and 1:10 antibody nanogel ratio (wt.); (e) targeting capability of ANCs at various antibody to nanogel ratios from 1:2 to 1:100 evaluated by flow cytometry.

We then focused on identifying the antibody density that would offer the best cellular targeting. Therefore, we evaluated a broad range of antibody to polymer ratios (from 1:100 to 1:2) for the conjugation reaction. After 4 hours of reaction, gel images show complete consumption of antibodies in all cases (Figure S30). We then evaluated the targeting capability of these ANCs using flow cytometry in BT474 cells. The MFI increases with increasing antibody surface density on the nanogel and the signal saturates between 1:2.75 to 1:2 antibody:polymer ratio (Figure 5e). To minimize off-target effects, it is necessary that we achieve maximum selectivity at minimal antibody density. These results suggest that the sweet-spot for the HER2 targeting in BT474 cells is ~1:2.75.

We used DiI as a model cargo to optically evaluate and optimize the ANCs. We were then interested in studying the relationship between ANCs targeting ability to the in vitro efficacy towards HER2+ cancer cells. We selected DM1, a microtubular inhibitor that has been widely used as ADC payloads, as the encapsulated drug for these studies. DM1 also offers to be covalently encapsulated to the polymer backbone of our nanogels through a thiol-disulfide exchange. DM1 was encapsulated in the three ANC scaffolds prepared with different conjugation chemistries. Concentration of the encapsulated DM1 was determined by LC-MS, and the DAR value was calculated based on the concentrations of dosed antibodies and drugs. We first evaluated the cytotoxicity of ANC^Tz-TCO-Lys at 1:2 1:5 and 1:10 antibody to polymer ratios to study the relationship between targeting ability and cytotoxicity (Figure 6a). Consistent with the DiI-based studies, cytotoxicity increased with increasing antibody density. Then we evaluated selective cell kill of ANC^N3-DBCO-Lys, ANC^Tz-TCO-Lys, and ANC^Tz-TCO-SS at the 1:2 antibody to polymer ratio in BT474 cells and MCF10A cells. We find that the cytotoxicity of ANCs via lysine-modified Ab (ANC^Tz-TCO-Lys) and site-specifically modified Ab (ANC^Tz-TCO-SS) based on the TCO-Tz reaction show similar cytotoxicity to BT474 cells, and both ANCs showed significantly higher cytotoxicity than ANC^N3-DBCO-Lys based on the DBCO-azide reaction (Figure 6b). No significant cytotoxicity of these ANCs was observed in MCF10A cells (Figure S34). Together, these results indicate that: (i) antibody decoration give rise to enhanced cytotoxicity in HER2+ cells but not in HER2- cells, i.e., cellular specificity; (ii) the antibody surface density influences cytotoxicity; and (iii) lysine-based modification does not significantly affect antibody orientation to affect cellular targeting

Figure 6.

(a) Cytotoxicity and EC₅₀ of DM1-ANCs at varied antibody to nanogel ratio in BT474 cells; (b) Cytotoxicity and EC₅₀ of DM1-ANCs prepared with three different conjugation chemistries; (c) preparation and characterization of homemade Trastuzumab-SMCC-DM1 ADC using LC-MS; (d) Cytotoxicity of ANC, ADC in BT474 cells dosed by DM1 concentration demonstrating superior cell killing efficiency with less antibody required; (e) Cytotoxicity of ANC, ADC, and Trastuzumab in BT474 cells dosed by antibody concentration; (f) Cytotoxicity of ADC, ANC in MCF10A cell; (g) Cytotoxicity of Trastuzumab in MCF10A cell, showing no significant toxicity in a HER2- cell line.

Comparison of ANCs-TDM1 vs ADC-TDM1.

Finally, we benchmarked our optimized ANC (ANC^Tz-TCO-Lys at 1:2 polymer antibody ratio) with a corresponding ADC, trastuzumab-SMCC-DM1 (TDM1). The ADC was simply prepared by conjugation of trastuzumab to SMCC-DM1, followed by purification by ultracentrifugation. The DAR of the resultant ADC is ~2.65, as characterized by LC-MS (Figure 6c). Note, the DAR characterization of ADC was performed without deglycosylation because of the uniform glycan structure of the obtained antibody (Figure S31, S32). The DAR of ANC was calculated to be 150, as described in the Materials and Methods section. Then we compared the cytotoxicity of ANC and ADC in both BT474 cells and MCF10A cells. As shown in Figure 6e, the optimized ANC exhibits significantly higher cytotoxicity compared to the ADC at the same antibody concentrations. This is attributed to ~56x higher DAR. Also, trastuzumab antibody itself is non-toxic to the cells. We also compared the ANC and the ADC based on DM1 concentration shown in Figure 6d, where the ADC was more efficacious than the ANC at low nanomolar concentrations, but the ANC showed significantly higher cytotoxicity than the ADC starting at DM1 concentration higher than 10 nM. The ANC, the ADC, and trastuzumab show no toxicity to the MCF10A cells even after 48 hours of incubation at the same concentrations (Figure 6f and 6g). Interestingly, cytotoxicity of TDM1 ADC plateaued at ~25% cytotoxicity which has been previously reported due to the limited cytosolic drug release in BT474 cells.⁵⁵ ANCs could potentially bypass this ADC recycle pathway leading to more efficient cell kill.

CONCLUSIONS

We outline a polymeric nanogel platform that allows for efficient surface functionalization with antibodies to generate ANCs, a versatile platform for targeted delivery. We focus on identifying and optimizing conjugation handles on the antibody and the biorthogonal reaction that offer high reaction efficiency and tunability. We show here that: (i) iEDDA-based TCO-Tz reaction offers a more robust conjugation method for attaching antibodies to nanogels, compared to the SPAAC-based DBCO-azide reaction. The high efficiency of the former reaction not only shortens the conjugation reaction time, but also obviates a post-conjugation purification step. (ii) site-specific conjugation achieved through a disulfide-rebridging method in antibodies offers similar efficiency as the more indiscriminate lysine-based conjugation method. We show that the advantage of the site-specific modification is relatively small in the Trastuzumab antibody studied here. However, considering that the reliability and efficiency of the disulfide-rebridging method is similar to that of the lysine-based conjugation, site-specific conjugation should be the preferred method for conjugation in general, because of the possibility of reactive lysines on the F_ab domain of other antibodies. (iii) density of antibodies on the nanogel surface can be tuned to optimize the avidity of the ANCs. It is reasonable to anticipate that there would be a balance between high avidity and cell specificity. The tunability in the density of the nanogels offers to enhance avidity, while concurrently identifying the sweet spot at which cellular specificity can be achieved. (iv) a direct comparison with the corresponding ADC highlights the advantages of the ANC platform. In addition to the avidity increase, ANCs offer to substantially increase the DAR. The increased DAR manifests itself in an increased cell kill efficiency by ANCs. Although the onset of targeted cell kill seems to be better for ADCs, the cell kill saturates at ~25% for the ADC. ANCs, on the other hand, offers substantially better cell kill and provides clear advantages in the amount of antibody needed. It is important to highlight that antibody-based treatments in cancer have resulted in the development of resistance. A recent study suggests the possibility of antibody exposure-dependent receptor degradation on cell surfaces.⁵⁶ Therefore, minimizing antibody exposure might be an advantageous feature in future therapies. Directing therapeutic reagents to cross biological barriers is still challenging in the field of nanomedicine. Biodistribution and tumor penetration are important parameters that significantly impact the outcome of tumor targeting therapeutics. As theses features in nanoparticle-based therapeutics could be significant influenced by the nanoparticle construct, the carrier platform must have sufficient tunability in their physiochemical properties such as size, surface characteristics, etc. for in vivo screening. Overall, the ANC platform outlined here bears these advantageous features and has the potential to make a translational impact.

Materials and Methods

All chemicals and reagents were purchased from commercial sources and were used as received unless otherwise mentioned. ¹H NMR spectra were recorded on a 400 MHz Bruker NMR spectrometer using the residual proton resonance of the solvent as the internal standard. Chemical shifts are reported in parts per million (ppm). ¹³C NMR spectra were proton decoupled and recorded on a 100 MHz NMR spectrometer using the carbon signal of the deuterated solvent as the internal standard. UV-Visible and fluorescence readings were obtained from Molecular Devices Spectramax iD5 plate reader. The small scale UV-vis reading were performed on BioDrop Touch Duo. Dynamic light scattering (DLS) measurements were performed using Malvern Zetasizer Nano ZS. Size Exclusion Chromatography (SEC) was performed on Agilent 1260 Infinity with a UV-Vis detector using a Cytiva Superdex 200 Increase 10/300 GL. Flow cytometry data were obtained from BD Dual LSRFortessa using FACSDiva software and analyzed on FlowJo. Mass spectrometry characterization of small molecules was carried out on Bruker MicroTOF ESI, while the characterization of antibody modification was performed on Thermo Orbitrap Fusion. SDS-PAGE gels were imaged by ChemiDocTM Imaging System and analyzed by Image Lab from Biorad.

Monomer and polymer synthesis:

Polymers P0 and P3 were synthesized according to previous procedures.²³ Briefly, a solution of RAFT Reagent (200 mg, 1 eq), PDSMA (94.58 mg, 10 eq), and AIBN (1.22mg, 0.2 eq) in THF (200 µL) was degassed by three freeze-pump-thaw cycles before being sealed off under argon atmosphere. After 6 hours at 65 °C, the polymerization media was concentrated under reduced pressure, diluted in DCM, precipitated in cold diethyl ether three times to remove unreacted monomers. The precipitate was collected and dried under vacuum. The resultant polymer was characterized by comparison of spectra with literature.¹⁴

Polymers P1, P2, and P4 were prepared in 2 steps: 1) Boc deprotection to obtain P0: 200 mg of Boc protected BCP was dissolved in a mixture of 1 mL DCM and 1 mL TFA and stirred for 5 minutes at 0 °C. The mixture was then warmed back to room temperature and stirred for 2 h. The solvent was then removed in vacuo, and the polymer was precipitated in cold diethyl ether (3x) to remove the remaining TFA. The precipitate was collected and dried under vacuum. 2) The deprotected polymer was dissolved in anhydrous dimethylformamide (DMF). Then, excess of triethylamine (TEA) was added, and the solution was stirred for 30 minutes. Then, 3 equivalents of azide-PEG4-NHS(P1), tetrazine-NHS (P2), or Cy7-NHS (P4) per amine were added, and the reaction mixture was stirred for 24 hours. The mixture was then dialyzed (Spectra/Por, ThermoFisher Scientific, MWCO = 3.5kDa) against a 4:1 mixture of dichloromethane (DCM): methanol for 2 days. The resultant polymer P1 was characterized by comparison of spectra with literature.¹⁴ Polymer P2 and P4 were characterized by ¹H and ¹³C NMR (Figure S2, S3, S19, S20).

Synthesis of Dibromo-Pyridazinediones Derivatives:

The synthesis of molecules 2–4 was carried out according to the previously reported protocols.⁵⁴ The structures of these molecules were confirmed by ¹H NMR (Figure S23–25).

Synthesis of molecule 5:

To a round-bottom flask, 91.1 mg (0.257 mmol) of 4 was dissolved in 2 mL of dry THF. 56.6 µL (0.515 mmol) of N-methylmorpholine was added and stirred at 0 °C for 10 minutes. After that, 40.1 µL (0.309 mmol) of isobutyl chloroformate was added to the reaction mixture and stirred at 0 °C for another 20 minutes. In a separate vial, TCO-PEG₂-NH₂ was dissolved in 1 mL of dry THF. The two reaction mixtures were combined and warmed up to room temperature with stirring overnight. The solvent was removed under reduced pressure, and the reaction mixture was purified with silica gel flash column (50–100% ethyl acetate in hexanes). Yield: 106.9 mg (65%); NMR (Figure S26, 27); ESI-MS m/z (5+Na+) = 661.04 (Calculated m/z = 661.07)

Antibody Modification Using Surface Lysine:

200 µg (39 µL) Trastuzumab (5.13 mg/mL used as reported from Selleckchem) in PBS buffer pH 7.4 were mixed with 1.0 µL of either DBCO-PEG5-NHS ester or TCO-PEG4-NHS ester (10 mM prepared in DMSO, 7.5 eq). The mixture was vortexed and incubated at 4 °C for 16 hours. The excess linker was washed with PBS and purified by Amicon ultra centrifugal kit (10 kDa cutoff) 6 times. The final concentration of antibodies was calibrated by BioDrop for further use.

Antibody Modification Using Disulfide:

200 µg (39 µL) Trastuzumab (5.13 mg/mL used as reported from Selleckchem) was diluted with 343 µL of the conjugation buffer (BBS pH = 8.0, 2% DMSO). 5.12 µL (10 eq) of molecule 5 (5 mM in DMSO) was added to the antibody solution and incubated at 4 ^oC for 60 minutes. After that, 2.68 µL (20 eq) of tris(2-carboxyethyl) phosphine (TCEP) (5 mM in 1x PBS) was added to the mixture and incubated at 4 °C overnight. The excess linker was washed with PBS and purified by Amicon ultra centrifugal kit (10 kDa cutoff) 6 times. The final concentration of antibodies was calibrated by BioDrop for further use.

Preparation of Trastuzumab-SMCC-DM1 ADC:

200 µg Trastuzumab (5.13 mg/mL used as reported from Selleckchem) was diluted with the conjugation buffer (100 mM PBS buffer pH 7.5, 150 mM NaCl) to a final concentration of 2 mg/mL. Then 15 eq of SMCC-DM1(obtained from MedChem Express) in 5% total solution volume was added to the solution and incubated at 4 ^oC for 24 hours. Then, the excess SMCC-DM1 was purified by Amicon ultra centrifugal kit (10 kDa cutoff) 6 times using PBS buffer. The final concentration of antibodies was calibrated by BioDrop for further use.

Characterization of Antibody Linker Conjugate and Antibody Drug Conjugate Using LC-MS:

For LAR analysis, antibody linker conjugates were deglycosylated before further characterizations by LC-MS (Thermal Fisher Orbitrap). To deglycosylate samples, we followed the established protocol using PNGase F from New England Biolabs. Briefly, 100 µg of the conjugate was mixed with 4 µL of 10x Glycol buffer and 0.5 µL of PNGase F. The volume of the mixture was adjusted to 40 µL by using HPLC water. The solution was then incubated at 37 °C for 24 hours. The resultant conjugate was then purified by SEC and concentrated to ≥1 mg/mL by Amicon ultra centrifugal kit (10 kDa cutoff).

The DAR of the Trastuzumab-SMCC-DM1 ADC was characterized directly without deglycosylation since the antibody used has shown one specific glycol side chain, results shown in Figure S31, S32.

Nanogel Formulation:

The polymeric nanogels were prepared according to a previously published procedure.¹⁴ A 1:1 (wt.) mixture of P1 or P2 (polymers with Ab-conjugation handles) and P3 (OMe-terminal polymer) was utilized to produce the nanogels. Briefly, P1, P2, and P3 were freshly prepared at 50 mg/mL in DMSO. DM1 and DiI were freshly prepared at a concentration of 10 mg/mL in DMSO. 20 µL of P1 (for N3 particles) or P2 (for Tz particles) and 20 µL of P3 were mixed with a 30 µL (15 wt%) DM1 or 20 µL (10 wt%) DiI stock solution. To this mixture of polymer and cargoes, HPLC water was used to adjust the volume to 1 mL, followed by immediate sonication for 5–10 minutes. DMSO was removed by dialysis against DI water using a mini dialysis kit (1 kDa cutoff) from Cytiva for 6 hours. After that, the nanogels were crosslinked using 7.7 µL of DTT (stock solution in water, 5 mg/mL) to target a 20% crosslinking density. After 3 hours of crosslinking, nanogels were further purified using a mini dialysis kit (8 kDa cutoff) from Cytiva to get rid of crosslinking byproducts and free drugs for 24 hours. The final polymer concentration of nanogels was calibrated to 2 mg/mL.

For the Cy7 NGs, a 5:4:1 (wt.) mixture of P1 or P2, P3, and P4 (Cy7 polymer) was used. P1, P2, P3, and P4 were freshly prepared at 50 mg/mL in DMSO, while DiI solution was prepared at a concentration of 10 mg/mL in DMSO. 20 µL of P1 (for N₃ particles) or P2 (for Tz particles), 16 µL of P3, and 4 µL of P4 were mixed with 20 µL (10 wt%) DiI stock solution. The sonication and purification processes were carried out similarly to the earlier procedure.

Antibody Nanogel Conjugation:

The antibody nanogel conjugation was carried out by incubating a mixture of 300 µL nanogel solution (0.6 mg polymer) and the desired amount of modified antibody for 24 hours at 4 °C for an efficient conjugation. The variation of conjugation degree was reported as a weight ratio between the antibody mass and polymer mass of the nanogel. For example, the ANC with the mass ratio of 1:2 was formulated with 0.3 mg of an antibody and 0.6 mg of the polymer input. This ratio translates to a molar ratio of 1:40 antibody-to-polymer.

Nanogel Size Characterization Using Dynamic Light Scattering: Dynamic light scattering (DLS) measurements were performed using Malvern Zetasizer Nano ZS with a 637 nm laser source with non-invasive backscattering technology detected at 173°. 70 µL of the sample was used each time. The sizes were measured after 2 minutes of temperature stabilization and reported as hydrodynamic diameters. The size measurement for all sample was performed in triplicate.

Second-order Kinetics Simulation:

The reaction progress was calculated based on second-order kinetics of two starting materials, rate = k[A][B], A = N₃ or Tz, B = DBCO or TCO. The rate law was derived according to the previously published work.⁵⁷ Based on our conjugation conditions, we translated the obtained LAR to the concentration of DBCO or TCO in the conjugation mixture to be around 50 μM. Similarly, the concentration of N₃ or Tz was calculated to be around 130 μM according to the polymer concentration formulated into the nanogels. We applied these to the obtained rate law to calculate the percent of reactant B consumed as a function of time shown in Figure 2A.

ANC DAR Determination:

The DAR of antibody nanogel conjugates used in the ADC ANC comparison study was calculated using the concentration of DM1 divided by the concentration of antibody. The concentration of DM1 encapsulated in the ANC was determined by LC-MS. Briefly, 10 µL of ANC solution was added to 90 µL of 100 mM TCEP in 50% MeCN/H₂O. The resultant solution was kept for 3 hours before LC-MS analysis. The standard curve of DM1 was generated using MS detector shown in Figure S35. The analysis was done in triplicates, and the average concentration was used to calculate the DAR of ANC. The concentration of the antibody was determined by BCA assay and calibrated according to the given concentration from purchase source. Since no further purification was needed for Tz-TCO ANCs, the amount of antibody is estimated to be the same as the dosed antibody for conjugation.

Sodium Dodecyl Sulfate–Polyacrylamide Gel Electrophoresis (SDS-PAGE) Gel Protocol:

The gel running protocol was adapted from the suggested protocol for Bolt Bis-Tris^™ 8% from ThermoFisher Scientific. Typically, each loading sample was calculated to be 1 µg of protein per well. To prepare 20 µL of gel loading sample, 5 µL of 4x LDS loading buffer was added to the desired amount of the sample. 1x PBS was then used for adjusting the volume to 20 µL. The sample was then denatured at 85 °C for 5 minutes. After the heated samples were vortexed and centrifuged to collect everything, the samples were loaded onto Bolt Bis-Tris 8% gel, along with the molecular weight ladder from PageRuler^™ Plus Prestained Protein Ladder. The samples were run with 1x MOPS SDS running buffer, prepared from 20x Bolt^™ MOPS SDS running buffer purchased from ThermoFisher Scientific, using 200 V for 30 minutes. The obtained gel was stained by Coomassie blue and imaged.

Size-Exclusion Chromatography (SEC):

Superdex 200 10/300 GL (purchased from GE) column was used for running the samples. The samples were run in 1x PBS buffer with 5 mM EDTA at 0.5 mL/min for 60 minutes. The presence of the antibody was monitored by using the absorbance at 280 nm. In case of ANCs, the obtained samples were concentrated using 10 kDa cut-off Amicon ultra centrifugal filter. Antibody quantification of ANCs before and after separation by SEC was performed by using Pierce^™ BCA protein assay kit following the recommended protocol with bovine gamma globulin (BGG) as a standard protein. The nanogel loading concentration for each ANC was calibrated by using Cy7 fluorescence.

Cell Culture:

All the cell lines used in this work were purchased from the American Type Culture Collection (ATCC). BT474 was maintained in DMEM:F12 (50:50) supplemented with 10% FBS, 100 units/mL of penicillin-streptomycin, and grown in 5% CO2 incubator. MCF10A cells were grown in DMEM: F12 media supplemented with 1% 1M HEPES, 1% of L-glutamine, 750 µL of 10 mg/mL gentamicin, 5% FBS, 500 µL of 10 mg/mL insulin, 1 mL of 10 μg/mL EGF (human), 100 µL of 500 μg/mL cholera enterotoxin and 5mL of 50 μg/mL hydrocortisone grown in 5% CO₂ incubator.

Flow Cytometry:

4 X 10⁵ of BT474 and MCF10A cells were treated with DiI-encapsulated nanogels or ANCs. The amount of nanogels or ANCs were calibrated according to the DiI concentration, which was monitored using UV-vis. Cells were incubated for 10 minutes. Then, the media was removed, and cells were washed twice with PBS. Cells were then trypsinized, quenched with media, transferred to centrifuge tubes, and spun at 2000 rpm for 5 minutes. Cells were resuspended twice in 200 µL PBS 1x. The final cell pallet was resuspended in 100 µL PBS 1x before being subjected to flow cytometry analysis. All studies were performed in triplicate.

For Cy7-labeled ANC targeting experiments, 1.5 X 10⁵ of BT474 and MCF10A cells in 200 µL complete media were plated in a 48-well tissue culture plate and incubated for 24 hours. The original media was removed, and the cells were then exposed to of nanogels or ANCs composed of 0.01 mg/mL Cy7-labeled polymers (0.1 mg/mL of total polymer) in 100 µL complete media. The concentration of the nanogels or ANCs were determined based on Cy7 fluorescence. After 10 minutes of treatment, the media was removed, and the cells were washed 3 times with 1x PBS. The cells were trypsinized, quenched with media, transferred to centrifuge tubes, spun at 2000 rpm for 5 minutes, and resuspended twice in 200 µL PBS 1x. The final cell pallet was resuspended in 100 µL PBS 1x before being subjected to flow cytometry analysis, monitoring the Cy7 and DiI channels. All studies were performed in triplicate.

Confocal Imaging:

1 × 10⁴ cells were first plated onto a 4-chamber 35 mm glass-bottom dishes and incubated at 37 °C in 5% CO₂ overnight. Then, cells were washed once with PBS before treated with new media containing 30 μg/mL DiI-encapsulated nanogels or ANCs at same DiI concentration (normalized by UV-vis). After incubating at 37 °C for 30 minutes, cells were washed twice with PBS and fresh media were added and continued incubating for 3.5 h. To prepare for visualization, the cells were washed with PBS and then stained with Hoechst 33342. Assessment of the nanogels or ANCs intracellular uptake was recorded using a 560 nm laser, and the nuclear stain was detected using a 405 nm wavelength laser. Imaging was performed on a Nikon Ti2 stand with a spinning disk confocal and 2 camera TIRF system. Co-localization of blue (Hoechst) and red (Cy5) channels were studied to evaluate the location of the dye-encapsulated particles.

Cell Viability Assay:

Cells were seeded on opaque flat-bottom 96-well tissue culture plates at a density of 5000 cells/well and rested for 24 hours at 37 °C in 5% CO₂. After incubation, the culture medium was removed, and cells were treated with fresh media containing nanogels or ANCs at different concentrations for 10 minutes. The drug-containing media was then removed and replaced with fresh media followed by incubation for 48 hours. After treatments, the cells were treated with Cell Titer-Glo reagent for 10 minutes at room temperature. The luminescent signal was measured using a SpectraMax iD5 multi-mode microplate reader. All studies were performed in triplicate.

ABBREVIATIONS

ADC

antibody drug conjugate

ANC

antibody nanogel conjugate

DAR

drug-to-antibody ratio

Tz

tetrazine

N₃

azide

TCO

trans-cyclooctene

DBCO

dibenzocyclooctyne

RAFT

reversible addition−fragmentation chain transfer polymerization

SEC

size exclusion chromatography

SMCC

succinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxylate