In Vitro Evaluation of Anti-HIV Radioimmunoconjugates Labeled with Astatine-211, Thorium-227 and Actinium-225

Abstract

We conducted an in vitro investigation of the selective cytotoxicity of alpha-emitting radioimmunoconjugates (α-RICs) directed against cells expressing HIV envelope (Env) proteins. It is well known that monoclonal antibody (mAb)-targeted α-emitting radionuclides can effectively kill antigen-expressing cells; however, the expected low-level expression of HIV antigens on latently infected cells poses an obstacle to all anti-HIV immune-based treatments, including α-RICs. This investigation tested the cytotoxicity of the HIV envelope antigen-binding mAbs, PGT126 (binding gp120) and 7B2 (binding gp41), conjugated with labeling chelators that bind the α-emitters astatine-211 (²¹¹At), actinium-225 (²²⁵Ac) or thorium-227 (²²⁷Th).

Methods:

High specific activity (SA) preparations of the α-RICs were made to increase the proportion of mAb conjugates carrying the α-emitting isotope. RIC cytolytic activity was evaluated against a cell line stably expressing the HIV envelope.

Results:

²¹¹At-labeled mAb conjugates did not demonstrate specific cell killing, while the longer lived radiometal α-RICs, ²²⁷Th and ²²⁵Ac, efficiently and specifically killed HIV envelope expressing cells.

Conclusions:

Potential explanations for these differential effects include the longer half-lives of ²²⁵Ac and ²²⁷Th compared to ²¹¹At and differences in the decay properties of radiometals compared to radiohalogens. These encouraging in vitro results suggest that in vivo evaluations of α-RIC in depleting the HIV harboring cells are warranted.

Graphical Abstract

1. INTRODUCTION

In 2022, the Centers Disease Control and Prevention estimated that there were 1.2 million people in the U.S. living with human immunodeficiency virus (HIV) inclusive of nearly 32,000 newly infected people [1]. Approximately three-quarters of these people were receiving anti-retroviral therapy leading to virologic suppression in 66%. In fact, combination antiretroviral therapy (cART) can suppress viral replication in the great majority of people who take it, leading to immune restoration, improved health and longevity, and the prevention of both vertical and horizontal transmission [2, 3]. However, viral suppression is not a cure and if treatment is interrupted, viral replication and disease progression resume [4-6]. Additionally, cART is costly, is not without toxicity, does not eliminate all the pro-inflammatory effects of HIV, and does nothing to alter the stigma of HIV. Thus, curing people of HIV by targeting and eliminating the cells that comprise the HIV reservoir remains a research priority [7].

Anti-HIV monoclonal antibodies (mAb) have been proposed and tested in strategies to deplete the HIV reservoir that persists despite effective antiretroviral therapy [6, 8-17]. Cytotoxic mAbs kill target cells by Fc-mediated complement fixation, antibody-dependent cellular cytotoxicity (ADCC), or antibody-dependent phagocytosis. However, not all anti-HIV mAbs elicit Fc-mediated effects or do so incompletely due to impaired cellular immune responses. As an alternative approach, we set out to develop radioimmunoconjugates (RIC) with enhanced cytotoxicity compared to unmodified mAbs. RICs are particularly effective reagents since they do not require internalization for cell killing, they can effectively kill resting cells, and they are less immunogenic than toxin- or drug-based immunoconjugates [18-20]. We chose to use α-emitting isotopes in the construction of the RICs given their exceptionally high linear energy transfer (conferring lethal cytotoxicity) combined with their short path-length that simultaneously minimizes off-target toxicity [21]. The high ionizing density of α-particles induces double-stranded DNA breaks that initiate a chain of events causing cell death; therefore, cytotoxicity is independent of cell cycle and dose rate [22]. The α-emitting radioimmunoconjugates (α-RICs) developed here contained the radionuclides astatine-211 (²¹¹At), thorium-227 (²²⁷Th) or actinium-225 (²²⁵Ac) conjugated to mAbs that target glycoprotein antigens within the HIV envelope (Env). In this investigation we evaluated the in vitro cytotoxicity of α-RICs made from one of two anti-HIV mAbs, PGT126 (targeting gp120) [23] and 7B2 (targeting gp41) [24, 25]. We used the previously characterized immunotoxin, 7B2, conjugated to de-glycosylated ricin A chain (7B2-dgA) [12, 25, 26] as the positive control in the cytotoxicity experiments. The target cells in these studies were HEK-293T (293T) cells that stably express the HIV Env HEK-293/92UG (92UG), parental HEK-293T cells that do not express HIV Env, and mixtures of 92UG and 293T cells [12, 27].

2. MATERIALS AND METHODS

2.1. General considerations

All reagents and solvents purchased from commercial sources were analytical grade or higher and were used without further purification unless noted. High purity nitric acid was obtained from Honeywell Fluka as TaceSELECT^™ Ultra grade. Metal-free water was obtained by demetallation performed in-house using Chelex 100 (BioRad). Solvents for high performance liquid chromatography (HPLC) were obtained as HPLC grade and filtered prior to use. Ligands used in this study were synthesized at the University of Washington (B10-NCS) or the University of British Columbia (pa ligands). The bifunctional ligands B10-NCS, H₄octapa-NCS, and H₄py4pa-NCS were prepared as previously published [28-31]

2.2. Radioactivity handling and measurement

All radioactive materials were handled according to approved protocols at the University of Washington, Seattle. All chelation reactions were conducted inside a radiochemical fume hood. Thorium-227 was received from Oak Ridge National Laboratory as a solid nitrate residue, freshly separated from its daughter ²²³Ra to reduce dose in the sample. Radiochemical yields were determined by instant thin layer chromatography (iTLC) and gamma spectroscopy, the latter being used to distinguish ²²⁷Th from its progeny. For ²²⁵Ac, iTLC was used to determine the radiochemical yield by allowing the developed iTLC strip to decay for 4 or more hours before counting on the gamma counter to remove the effects of its progeny. The amount of radioactivity (²²⁵Ac or ²²⁷Th) in the final samples of radiolabeled antibody was confirmed by gamma spectroscopy.

2.3. HPLC equipment and conditions

All reactions were monitored by HPLC. Samples were assessed on a system that contained a Hewlett–Packard 1050 isocratic pump, a 1050 Hewlett–Packard variable wavelength UV detector (280 nm), and a Beckman 170 gamma detector. Analyses of the HPLC data were conducted on Hewlett–Packard HPLC ChemStation software. Size exclusion HPLC (SE-HPLC) was carried out on BioRad 250 or a BioRad Select.

2.4. Antibody conjugation and radiolabeling

2.4.1. Ricin A chain immunotoxin preparation

The de-glycosylated ricin A chain conjugate, 7B2-dgA, was prepared by conjugation of dgA [32] (a gift from Ellen Vitetta (University of Texas Southwestern) to 7B2 using the heterobifunctional cross-linking agent succinimidyl 6-[3(2-pyridyldithio)propionamido]hexanoate (SPDP, Pierce Biotechnology, Rockford, IL USA) at a 1:1 mass ratio as described previously [26]. Products were characterized biochemically by SDS-PAGE and contained a mixture of 0, 1, 2, or 3 dgA per 7B2 molecule.

2.4.2. Radioimmunoconjugate preparation

The mAb conjugates PGT126-B10 and 7B2-B10, used for labeling with ²¹¹At, were prepared by conjugations of PGT126 or 7B2 with the bifunctional ligand B10-NCS using a previously described method [29]. The conjugate PGT126-octapa, used for labeling with ²²⁷Th, was prepared by conjugation of PGT126 with H₄octapa-NCS as previously described [33]. The mAb conjugate 7B2-py4pa, for labeling with ²²⁵Ac, was prepared by conjugation of 7B2 with H₄py4pa-NCS under the same reaction conditions. Prior to conjugation with the bifunctional picolinic acid reagents H₄octapa-NCS and H₄py4pa-NCS, the mAbs were concentrated to approximately 5 mg/mL and extensively de-metallated by dialysis against buffers that had been treated by passing over a Chelex column to remove trace metals [34]. De-metallation was conducted a second time after preparation of the mAb conjugates. After the final dialysis, the mAb conjugates were transferred to acid-washed plastic tubes for storage. Purity of the conjugates was assessed by SE-HPLC (see supplemental Figure S2 for PGT-126-octapa and 7B2-py4pa as examples). Information on the number of conjugates per mAb molecule was obtained by mass spectral analyses (see supplemental Figure S3, for 7B2-B10, as an example).

2.4.3. Antibody radiolabeling

The α-emitting radionuclides used were obtained as follows: ²¹¹At (t_1/2 = 7.21h) was obtained by irradiation of an alpha beam on a bismuth target [35], then isolated using a wet chemistry method [36]. The ²²⁷Th (t_1/2 = 18.7d) was obtained from the DOE Isotope Program’s National Isotope Development Center (NIDC) as a solid nitrate salt, which was treated with an oxidant to assure it was in the +4 state and purified from daughter radionuclides just prior to its use [28]. The ²²⁵Ac (t_1/2 = 9.92d) was obtained from the NIDC as a solid nitrate salt and was used as obtained. The PGT126-B10 and 7B2-B10 conjugates were radiolabeled with ²¹¹At without the addition of the chloramine-T oxidant using the labeling conditions previously described [29]. Because target cells are expected to express low levels of HIV Env, labeling reactions were conducted using small quantities of mAbs (100-300 μg) to obtain high SA-labeled products that maximized the percent of mAbs carrying an isotope (Table 1). Radiolabeling conditions were optimized to achieve high SAs for all α -RICs before purifying by size exclusion chromatography. Radio-SE-HPLC was used to determine the purity of the final product for ²¹¹At RICs and ²²⁵Ac in one experiment when the yield was sufficiently high (supplemental Figure S4). iTLC was used to determine the purity of the ²²⁷Th (91%) and ²²⁵Ac (97-99%) final products.

Table 1.

Summary of radiolabeling results for RIC preparations used in the experiments presented.

Radionuclide	Conjugate	Radiochemical yield	Specific activity (MBq/mg )*	Specific activity (mCi/mg )*	Antibody cencentration (mg/mL)	Activity concentration (MBq/mL)*	Activity concentration (MCi/mL)*	Ascorbic acid 225 mg/mL	% Antibody with Radionuclide*	# of Antibody per Radionuclide *
At-211	7B2-B10	50.5	902.8	24.4	0.084	75.85	2.05	Yes	0.84%	119
At-211	PGT126-B10	69.5	1198.8	32.4	0.108	129.87	3.51	Yes	1.12%	89
Th-227	PGT126-Octapa	13.4	36.63	0.99	0.044	1.591	0.043	Yes	2.13%	47
Ac-225	7B2-Py4pa	68.7	26.778	0.724	0.063	3.071	0.083	Yes	0.83%	121

^*At the end of radiolabeling.

2.4.3.1. ²¹¹At labeling

Sodium phosphate solution (0.5 M, pH 6.8) was combined with either 7B2-B10 or PGT126-B10 before adding [²¹¹At]NaAt (6.3–13.2 mCi, pH 6–6.5) with no oxidant. After 2 min at room temperature, the reaction was quenched by adding sodium metabisulfite. This was then purified on a PD-10 column and collected in PBS with ascorbic acid, USP (Ascor, McGuff Pharmaceuticals, Inc, Santa Ana, CA) added to a final concentration of 25 mg/mL as a radioprotective antioxidant.

2.4.3.2. ²²⁷Th labeling

De-metallated sodium citrate solution (0.05 M, pH 5.5) containing 1 mM EDTA was combined with 60 μL ²²⁷Th in 0.1 M HNO₃ (185 μCi). To this, 90 μL of 1 M de-metallated sodium citrate was added to adjust the pH to 5.5, followed by the addition of PGT126-octapa. This mixture was allowed to react for 2h at room temperature. For all the antibody labeling reactions, the product mixture was passed over a PD-10 column to separate the radiolabeled antibody from free radionuclides. The products were collected in PBS with ascorbic acid, USP (Ascor, McGuff Pharmaceuticals, Inc, Santa Ana, CA) added to a final concentration of 25 mg/mL.

2.4.3.3. ²²⁵Ac labeling

De-metallated HEPES buffer (0.05 M, pH 7) was combined with 7B2-py4pa. To this, 25-35 μL of ²²⁵Ac (212-430 μCi) in 0.1 M HNO₃ was added. This mixture was incubated for 1h at 37°C and then purified on a PD-10 column and collected in PBS with ascorbic acid, USP (Ascor, McGuff Pharmaceuticals, Inc, Santa Ana, CA) added to a final concentration of 25 mg/mL.

2.5. Cell lines and antibodies

HEK-293T (hereafter 293T) cells are a cell line derived from human embryonic kidney cells that have been modified to express the simian virus 40 (SV40) large T antigen and were obtained from ATCC. HEK-293T/92UG cells (hereafter 92UG) constitutively express Env of the Clade A clinical isolate 92UG037.8 in native conformation. The cells were the gift of Dr. Bing Chen, Boston Children’s Hospital [37]. Expression of Env on the surface of these cells was maintained by periodic selection with puromycin (1 μg/mL) and verified by flow cytometry with mAb VRC01 and fluorescent anti-human IgG as described previously [27, 38]. All cell lines were cultured at 37°C in a humidified CO₂ atmosphere and maintained by serial passage not to exceed a concentration of 10⁶ cell/mL. 293T and 92UG cells were grown in DMEM (GIBCO, 4.5 gm glucose/L) with 10% fetal bovine serum (Hyclone, Logan UT USA), and 1% penicillin/streptomycin (GIBCO).

MAb 7B2 binds CSGKLIC, the conserved 7 amino acid loop within the helix-loop-helix of gp41 [24], and has been found to be the most effective mAb for delivering anti-HIV immunotoxins [25, 27], which differ from RICs in that they must be internalized to kill. 7B2 was originally produced by Dr. James Robinson (Tulane University) and was the gift of Dr. Barton Haynes (Duke University). MAb PGT126 is a broadly neutralizing antibody targeting the base of the V3 glycan loop of gp120. It was produced as previously described [12].

2.6. HIV envelope flow cytometry staining

Following trypsinization, 2×10⁵ cells from each cell line were taken from culture for flow cytometry surface staining of the HIV envelope. Cells were incubated with 1 mL of 20% FBS-PBS for 10 min at 4°C to block the FcR. Cells were washed once with 1 mL 1X PBS and incubated with 2μg/100μl of either a HIV envelope binding antibody (Anti-HIV-1 gp120 Monoclonal VRC01 (NIH AIDS Reagent Program #12033)) or a primary human IgG1 kappa isotype control antibody (Sigma #I5154) for 25 min at room temperature. Following a second PBS wash, cells were incubated with 2μg/100μl of either a secondary PE Goat anti-Human IgG Fc antibody (Invitrogen 12-4998-82) or a secondary PE Goat IgG Isotype Ctrl antibody (Biolegend #403004) at for 25 min at room temperature. Cells were washed a final time with PBS, pelleted and run on a BD Celesta or Fortessa flow cytometer, and analyzed using FlowJo software.

2.7. ELISA for mAb-conjugate ligand binding

For the 7B2 ELISA, 10 ng of 7B2 peptide (sequence LGIWGCSGKLICTT) [39] in 100 μL PBS was incubated on Immulon 2HB plate (Thermo, Waltham, MA, USA) overnight at 4°C, followed by blocking in 200 μl 1% PBS-BSA blocking solution and incubation for 6–8h at room temperature or overnight at 4°C. Samples and standards were serially diluted in a U-bottom 96-well plate in blocking buffer. 100 μl of the dilutions were transferred to the Immulon plate and incubated for 6–8h at room temperature or overnight at 4°C, followed by 6 washes in 0.1% PBS-Tween-20 wash solution. 50 μl of secondary antibody (Sigma #A8792) diluted at 1:80,000 dilution in blocking buffer was added and incubated for 1h at room temperature, followed by 4 washes in wash buffer. The final steps include incubation with 200 μl of 3,3',5,5'tetramethyl-benzidine (TMB) solution (Sigma-Aldrich, St. Louis, MO) for 20 min on an orbital shaker at room temperature. The reaction was stopped by addition of 1 N H₂SO₄. The absorbance was read at 450nm using a BioTek EL800 microplate absorbance reader (Biotek, Winooski, VT). For the gp120 ELISA, 5 μg of concanavalin A per well of a 96-well maxisorp plate (Thermo Fisher Scientific, Waltham, MA) was incubated overnight at room temperature, followed by two washes with 0.05% PBS-Tween 20. The plates were then coated with 50 ng of HIV-1 BaL gp120 recombinant protein (NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH) for 1h at room temperature, washed four times, then incubated in blocking buffer (5% milk-0.05% Tween 20 in PBS) for 1h at room temperature. Samples and standards were serially diluted in a U bottom 96-well plate in blocking buffer, transferred to the maxisorp plate, incubated for 1h at room temperature, and washed four times before the addition of anti-human peroxidase-conjugated rabbit anti-human immunoglobulin G (IgG) secondary antibody (Sigma-Aldrich A8792, 1:80,000 dilution) for 1h at room temperature. The plates were washed four times before addition of the 3,3',5,5' tetramethyl-benzidine (TMB) solution (Sigma-Aldrich, St. Louis, MO) and incubated for 20 min at room temperature on an orbital shaker. The reaction was stopped by addition of 1 N H2SO4. The absorbance was read at 450nm using a BioTek EL800 microplate absorbance reader (Biotek, Winooski, VT). Standard curves for antibody quantification were established using purified monoclonal antibodies (anti- HIV-1 gp120 monoclonal VRC01, from Dr. John Mascola, NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH, and anti-HIV-1 gp120 monoclonal PGT128, from Dr. Denis Burton, IAVI Consortium). Supplemental Figures S5, S6 and S7 demonstrate the preserved binding of 7B2-B10 (with and without ²¹¹At), PGT126-B10 and PGT126-octopa to their respective ligands.

2.8. Cellular cytotoxicity assay

Cellular toxicity was measured on an xCELLigence Real-Time Cell Analysis (RTCA) DP instrument. This instrument measures the electrical impedance, by living, adherent cells, of a current passed across wells in specialized plates; the higher the impedance, the more confluent and intact the monolayer. This system allows for the continuous monitoring of the cell index (cell viability) over time. In these experiments, one day following plating, cell lines were treated with either Triton X-100, the immunotoxins 7B2-dgA or PGT126-dgA, cold RICs, or the different anti-HIV radiolabeled RICs at various radioactivity levels (see figure keys for final concentrations and specific activities). Each condition was performed in triplicate on 293T cells, plated at 2×10⁴ cells/well, and 92UG cells, plated at 1×10⁴ cells/well, in 150 μl of culture media in E-plate 16 PET (Agilent Technologies Cat#300600890). Soluble CD4-IgG2 (sCD4), which enhances 7B2-Env binding [12, 27], was added to cells treated with 7B2-based immunotoxins or RICs. Pre-dilution of the antibodies, immunoconjugates, sCD4 or triton X-100 were freshly prepared in cell culture media. 50 μl of the pre-dilution was added to each well already containing 150 μl of cell culture to reach the final antibody concentrations and radioactivity per well provided in the captions to each figure. Based on previous titrations with fluorescent mAbs, the antibody concentrations used in these experiments saturated cell-surface receptor binding sites. Non-bound RICs were removed by rinsing the cells once with 1X PBS after 1h incubation, followed by incubation at 37°C in cell culture media until the end of the experiment. The xCelligence RTCA DP software was set to measure cellular impedance (cell viability) every 15 min over 7 days after α-radiation. At the end of the experiment, the data were exported in Microsoft Excel and analyzed using Prism software.

2.9. Statistical analysis

Unless otherwise stated, all samples were run in triplicate. Results are shown as mean and SEM. If no error bars are visible, then the symbol on the graph is larger than the error. Calculations and statistical analyses were performed and graphs drawn using GraphPad Prism v9 (GraphPad, Boston, MA USA).

3. RESULTS

The goal of this investigation was to prepare four anti-HIV α-RICs and assess their cytotoxic function in vitro against a previously validated cell line expressing the Envelope of HIV-1 (Env) [37]. Expression of HIV Env on the surface of 92UG cells and the lack of that expression in the parental 293T cells was confirmed by flow cytometry. The results are shown in Figure 1 and confirm that more than 90% of the 92UG cells expressed the HIV Env.

Figure 1.

Flow cytometry measurement of HIV Env expression at the surface of parental 293T cells or 92UG cells stably expressing the HIV envelope. The cells were either unstained, stained with the secondary antibody only as a control, or stained with a primary VRC01 antibody directed at the HIV Env followed by staining with an anti-human secondary antibody conjugated to a PE fluorochrome. Side scatter is shown on the vertical axis and on the horizontal axis either forward scatter (left of each pair) or log fluorescent intensity (right). The percentage of cells falling within the indicated gate is indicated on the right side of each plot.

3.1. Ligands, antibody conjugation, and radiolabeling

Antibody labeling with ²¹¹At was performed by conjugation of a bifunctional boron cage (B10-NCS), while conjugation of the bifunctional chelating molecules H₄octapa-NCS and H₄py4pa-NCS mediated ²²⁵Ac and ²²⁷Th labelings, respectively (supplemental Figure S1). The number of isotope chelating molecules attached per mAb molecule can affect antigen binding with target cells. To test the effect of the chelating molecules on mAb affinity for HIV Env, we prepared mAb-conjugates with varying numbers of chelating moieties attached and used an ELISA to measure the binding efficiency to Env (gp120). Supplemental Figures S5, S6 and S7 demonstrate minimal effects of conjugate binding on the affinity of the mAbs for HIV Env. Supplemental Figure S5 also demonstrates that ascorbic acid preserves the integrity and binding of the RIC from degradation by the nuclide (in this case ²¹¹At).

3.2. Cytotoxicity, specificity, and kinetics of cell killing by α-RICs

The cytotoxic effects of the α-RICs were evaluated using the xCELLigence RTCA in an assay that quantifies cell death kinetics at regular intervals [40-45]. In these experiments, cells exposed to 1% non-ionic surfactant Triton X-100 demonstrated 100% non-specific killing; 92UG cells exposed to 7B2-dgA served as positive controls for Env-mediated cell killing, 92UG cells exposed to unlabeled mAb conjugates served as negative controls, and parental (HIV Env negative) 293T cells exposed to 7B2-dgA and the RICs provided data on the specificity of cell killing.

As shown in Figure 2A-B, the highest concentration of ²¹¹At-7B2-B10 (SA 740 MBq/mg; 20.00 mCi/mg) in the presence of sCD4, which enhances the binding of 7B2 to Env (gp41) [12, 27], had no or possibly a small, enhanced effect on cell growth at early time points (~48h) and caused a small and non-specific cytotoxic effect at later time points with no difference between Env-expressing (92UG) and non-expressing (293T) cells. The area under the curve (AUC) calculations, a quantitative measure of cell viability (Figure 2C), demonstrate no significant effect of ²¹¹At-7B2-B10 on 293T (Env-) cells, a small, enhanced growth effect on 92UG (Env+) cells at the lowest concentration of ²¹¹At, and no effect at higher concentrations of ²¹¹At. Similarly, ²¹¹At-PGT126-B10, at a SA of 1036 MBq/mg (28.00 mCi/mg), had a late cytotoxic effect on both 92UG and 293T cells (Figures 3A-C)), demonstrating only non-specific and incomplete killing.

Figure 2.

Cell index curves demonstrating cytotoxicity of ²¹¹At-7B2-B10 radioimmunoconjugate on 293T (A) and 92UG (B) cell lines. The RIC were added on cells 24 hours after plating the cells, then rinsed with 1X PBS one hour later. The cells were kept in culture for 96 hours, and the xCelligence RTCA device measured the cell index in each well every 15 minutes. Each condition was performed in triplicate. Means without SEM are plotted. Area Under the Curve (AUC) comparisons (C) for each treatment condition were calculated following the addition of the immunoconjugates in the culture. Unpaired t-tests were performed on triplicate AUC values in GraphPad Prism: * p-value<0.05; **<0.01; ***<0.001. This experiment is representative of three independent experiments.

Figure 3.

Cell index curves demonstrating cytotoxicity of ²¹¹At-PGT126-B10 on 293T (A) and 92UG (B) cell lines. The RIC were added on cells 24 hours after plating the cells, then rinsed with 1X PBS one hour later. The cells were kept in culture for 96 hours, and the xCelligence RTCA device measured the cell index in each well every 15 minutes. Means without SEM are plotted. Areas under the curves (AUC) comparisons (C) for each treatment condition were calculated following the addition of the immunoconjugates in the culture. Unpaired t-tests were performed on triplicate AUC values in GraphdPad Prism: * p-value<0.05; **<0.01; ***<0.001. This experiment is representative of two independent experiments.

3.3. Cytotoxicity, specificity, and kinetics of cell killing by ²²⁷Th-PGT126-octapa

Because isotopes with longer half-lives than ²¹¹At (7.2h) facilitate the construction of RICs with lower overall specific activities but comparable proportions of antibodies that carry a nuclide (see Table 1), we investigated the cytotoxic effect of a thorium-227 (half-life 18.7d)-based RIC, ²²⁷Th-PGT126-octapa, on Env-expressing and non-expressing cells. First, using an ELISA assay, we confirmed that conjugation of the PGT126 antibody to the octapa chelating conjugate did not impair its overall binding to the gp120 antigen (supplemental Figure S7).

Dose-dependent killing of ²²⁷Th-PGT126-octapa on 92UG Env-expressing cells was observed and complete at 96h by the two RIC preparations at the highest concentrations (18.5kBq and 37kBq/well or 0.5 and 1μCi/well) (Fig. 4B and C), while the impact on 293T cells was more moderate (Figure 4A and C). At lower concentrations (3.7kBq/well or 0.1μCi/well), there were non-specific effects on cell viability that were observed both on non-Env-expressing 293T and Env-expressing 92UG cell lines. At very late time points, both cell lines under control conditions (untreated and PGT126-octopa) died due to overcrowding.

Figure 4.

Cell index curves demonstrating cytotoxicity of ²²⁷Th-PGT126-octopa on 293T (A) and 92UG (B) cell lines. The RIC were added on cells 24 hours after plating the cells, then rinsed with 1X PBS one hour later. The cells were kept in culture for 168 hours, and the xCelligence RTCA device measured the cell index in each well every 15 minutes. Means without SEM are plotted. Areas under the curve (AUC) comparisons (C) for each treatment condition for the first 100 hours of the experiment (before cell outgrowth and death due to overcrowding occurred). AUCs were calculated for each condition following the addition of the immunoconjugates in the culture. Unpaired t-tests were performed on triplicate AUC values in GraphPad Prism: * p-value<0.05; **<0.01; ***<0.001.

3.4. Cytotoxicity, specificity, kinetics, and bystander killing by ²²⁵Ac-7B2-py4pa

Because of the efficacy of ²²⁷Th-PGT126-octapa and because additional ²²⁷Th was no longer available, we developed an actinium-225 (²²⁵Ac, half-life 9.92d) RIC and tested the cytotoxic effect of the ²²⁵Ac-7B2-py4pa on Env-expressing and non-expressing cells. Potent and specific killing of the 92UG Env-expressing cell line by ²²⁵Ac-7B2-py4pa was demonstrated (Figure 5B-C). In contrast, there was no effect of ²²⁵Ac-7B2-py4pa on non-Env expressing 293T cells relative to untreated cells (Figure 5A and C).

Figure 5.

Cell index curves demonstrating cytotoxicity of ²²⁵Ac-7B2-py4pa on 293T (A) and 92UG (B) cell lines. The RIC were added on cells 24 hours after plating the cells, then rinsed with 1X PBS one hour later. The cells were kept in culture for 190 hours, and the xCelligence RTCA device measured the cell index in each well every 15 minutes. Means without SEM are plotted. Areas under the curve (AUC) comparisons (C) for each treatment condition for the first 100 hours of the experiment (before cell outgrowth and death due to overcrowding occurred). AUCs were calculated for each condition following the addition of the immunoconjugates in the culture. Unpaired t-tests were performed on triplicate AUC values in GraphPad Prism: * p-value<0.05; **<0.01; ***<0.001.

Lastly, a cell mixing experiment was conducted to assess the off-target cytotoxicity of ²²⁵Ac-7B2-py4pa. 293T and 92UG cells were co-plated at a ratio of 1:10, respectively, and exposed to ²²⁵Ac-7B2-py4pa. After 3 days there was a clear cytotoxic effect on the mixed cell culture, and by 6-7 days the monolayer was destroyed without evidence of outgrowth of the non-Env expressing 293T cells, suggesting an off-target effect on 293T cells (Figure 6).

Figure 6.

Cell index curves demonstrating cytotoxicity of ²²⁵Ac-7B2-py4pa on a mixed culture of non-Env expressing 293T and Env-expressing 92UG cell lines at a 1:10 ratio (293T:92UG) (A). The RIC were added on the cells 24 hours after plating the cells, then rinsed with 1X PBS one hour later. The cells were kept in culture for 7 days, and the xCelligence RTCA device measured the cell index in each well every 15 minutes. Means without SEM are plotted. Areas under the curve (AUC) comparisons (B) for each treatment condition for the first 100 hours of the experiment (before cell outgrowth and death due to overcrowding occurred). AUCs were calculated for each condition following the addition of the immunoconjugates in the culture. Unpaired t-tests were performed on triplicate AUC values in GraphPad Prism: * p-value<0.05.

4. DISCUSSION

Here we describe an investigation of α-RICs containing one of three different α-emitting nuclides, ²¹¹At, ²²⁷Th, and ²²⁵Ac, to evaluate their cytotoxicity on cells expressing HIV Env. We chose to develop anti-HIV RICs because, unlike other treatment options [7], they work independently of the immune system; they do not require intact host immunity to target and eliminate the HIV harboring cells [46]. Furthermore, RICs kill both dividing and non-dividing cells, only cell-surface binding is necessary for eradication, and they are not immunogenic, unlike immunotoxins [19, 26] and possibly drug conjugates [47, 48]. α-RICs offer the additional advantage that they target and selectively kill individual cells or small clumps of cells due to their high linear energy transfer (LET ~ 100 keV/μm), which inflicts lethal dsDNA breaks that overwhelm repair mechanisms. This effect is independent of dose rate, oxygen tension, or radiosensitivity that are characteristics of β-emitting RICs. Further, the short pathlength of α-emitters limits killing to targeted and immediately neighboring cells, substantially reducing unwanted organ and tissue toxicity [21]. To date, most clinical trials using anti-HIV mAbs have employed broadly neutralizing antibodies (bNAbs), reasoning that their neutralization potential is key to their effectiveness. However, neutralizing antibodies frequently select for viral escape variants [49, 50] that can rapidly dimmish their effectiveness. In contrast, the neutralization potential (and therefore the probability of viral escape) of mAbs used to construct immunotoxins (IT) is not a determinant of their effectiveness [25]. While this remains to be investigated for RICs, it is plausible that the use of non-neutralizing antibodies, such as 7B2, to construct RICs will present a lower risk for antibody-driven viral escape. Additionally, combinations of neutralizing and/or non-neutralizing mAbs could be used to generate RICs targeting various Env epitopes.

RICs have been used extensively in humans to treat cancer and as conditioning treatments for stem cell transplantation, demonstrating their effectiveness and safety [51-53]. Recently, clinical trials have been initiated using RICs labeled with ²¹¹At [54, 55], ²²⁷Th and ²²⁵Ac [56-60], highlighting the potential of these radionuclides for clinical applications. RICs are particularly suitable reagents for depleting the HIV reservoir since they can eliminate rare, widely dispersed cells with limited Env expression such as those that constitute the reservoir.

Other investigators have explored the efficacy of different anti-HIV RICs using mAb conjugates of 2556 and 246-D, which are directed against gp41 [18, 19]. Those conjugates were labeled with the α-emitting radionuclides ²¹³Bi and ²²⁵Ac, and the β-emitting radionuclides rhenium-188 (¹⁸⁸Re) or lutetium-177 (¹⁷⁷Lu) [18, 19, 61, 62]. The investigators demonstrated effective in vitro killing of chronically infected T cells (ACH-2) and HIV-infected human peripheral blood mononuclear cells (hPBMCs), as well as in vivo cytotoxicity against HIV-infected hPBMCs implanted into SCID mice [18, 19]. Importantly, the target cells used in these experiments were either chronically infected T-cell lines or human PBMCs that were productively infected with HIV. These cells expressed high numbers of mAb binding sites on their cell surface (calculated to be >10⁵–10⁶ per cell) [19], which is more than enough to capture sufficient numbers of RICs for efficient cell killing. In one of these studies, ex vivo elimination of PBMCs harboring replication competent HIV from persons with HIV on anti-retroviral therapy using a ²¹³Bi-RIC was demonstrated. However, many subjects in these experiments had detectable plasma HIV RNA, there was no correlation between plasma HIV RNA levels and the susceptibility of PBMCs to the RIC, and direct measurements of PBMC Env expression at the time of RIC exposure was not provided [61]. In vivo, and especially in patients on suppressive antiretroviral therapy (ART), Env expression is tightly regulated by storage in CTLA-4-containing secretory granules until late in the viral cycle, thereby limiting immune system exposure and T-cell activation [63-65]. HIV virions and the cells that produce them express low levels of the HIV Env (estimated ~7-14 Env trimers/virion) [66], posing a challenge for both humoral and cellular immune responses as well as immunotherapies that rely on surface ligands to recognize and clear infected cells. In one human study, positron emission tomography (PET) using VRC01 conjugated to Zirconium89 (⁸⁹Zr-VRC01) did detect HIV Env in tissues in persons with HIV who were viremic or on suppressive ART. However, the differences between persons with HIV and uninfected control patients were most prominent for those who were viremic. Env was detected in patients with HIV on suppressive ART, but not in all tissues and not by all measurements. Significant levels of Env were detected in bone marrow, inguinal (but not axillary) lymph nodes, intestine and nasal turbinates but not in other tissues [67]. In similar experiments using macaques infected with simian immunodeficiency virus (SIV), SIV Env was detected by PET in animals on suppressive ART for 5 weeks and in elite controllers (animals that are not viremic without ART) demonstrating that Env expression, and possibly viral replication, can be detected in aviremic animals using this SIV model [68]. Any treatments that target the latent reservoir must contend with the challenge that latently infected cells, even those induced to express HIV Env, will likely have low Env surface expression. Similarly, low Env expression levels may also be an obstacle for anti-HIV α-RICs since the high SA required to effectively target these cells can be difficult to obtain and can result in degradation of the mAb conjugate due to radiolysis as observed with ²¹¹At-7B2 RIC (supplemental Figure S5, ²¹¹At-7B2 RIC without ascorbic acid). To address this challenge, RICs could be used in combination with other treatment options such as latency-reversing agents (LRAs) in a “shock and kill” approach that would increase the availability of targeted Env antigen.

It is estimated that efficient cell killing by α-RIC may require the binding of 5–10 RICs/cell since emitted α-particles can decay in any direction and only a fraction follow a path through the cell nucleus, disrupting dsDNA leading to cell death [69]. The low number of Env molecules expressed on HIV infected cells in vivo is potentially problematic since radiolabeled mAb preparations generally contain many more unlabeled mAb molecules than those carrying a radionuclide. Within an RIC preparation, the proportion of mAb molecules that contain a radionuclide can be determined by the SA (e.g., MBq/mg) and half-life of the radionuclide such that isotopes with shorter half-lives decay more rapidly (release more α-particles/atom/unit of time), resulting in fewer radionuclide atoms per mAb molecule for any given level of SA (Table 1).

Our initial studies focused on ²¹¹At-labeled mAbs since ²¹¹At is readily available at our facility and is being evaluated in clinical trials of blood-borne cancers (NCT03670966, NCT04083183, NCT03128034, NCT04466475) [55]. The majority of those studies involve targeting the CD45 antigen on hematopoietic cells that typically have ~200,000 antigens/cell, making RICs with low SAs adequate. For those studies, SAs in the range of 15–30MBq/mg (0.41-0.81mCi/mg) provide effective hematopoietic cell killing. Given the expected low Env expression levels encountered in persons on suppressive ART, it was clear that much higher SAs would be required to effectively target these cells. Therefore, the ²¹¹At-based RICs used in our early investigations, including the experiments presented here, were prepared by reacting a large amount of ²¹¹At (232–490MBq; 6.27–13.2mCi/mg) with small quantities (0.1–0.3mg) of the mAb-B10 to yield SAs from 451–1841MBq/mg (12.2–49.8mCi/mg) for ²¹¹At-7B2-B10 and 798–1199MBq/mg (21.6-32.4mCi/mg) for ²¹¹At-PGT126-B10. Although we demonstrated that conjugated and nuclide-labeled mAbs retained their affinity for HIV Env (Supplemental Figure S5), none of the ²¹¹At-mAb-B10 conjugates effectively and specifically killed the 92UG target cells (Figures 2 and 3).

Since we had previously demonstrated that an α-emitting RIC labeled with ²²⁷Th was effective in treatment of hepatocellular cancer in an orthotopic murine model [33], we next chose to develop an anti-HIV RIC using ²²⁷Th. We found the labeling of a PGT126-octapa conjugate with ²²⁷Th surprisingly difficult compared to our previous experience, providing only a low radiochemical yield (i.e., 13%) of ²²⁷Th-PGT126-octapa. Nevertheless, the isolated ²²⁷Th-PGT126-octapa demonstrated specific and dose-dependent cell killing (Figure 4). After this initial ²²⁷Th allocation, additional shipments were not available during our funded study period. However, since the ²²⁷Th-PGT126-octapa results were encouraging, we decided to evaluate the α-emitter ²²⁵Ac since its half-life is similar to that of ²²⁷Th. In several experiments we prepared and tested ²²⁵Ac-7B2-py4pa with SAs from 15–68MBq/mg (0.41–1.8mCi/mg) and, again, demonstrated dose-dependent and specific killing of HIV Env expressing cells as shown in Figure 5, representing one of these experiments. We did observe enhanced cell growth early on after exposure of both 92UG and parental 293T cells exposed to low doses of ²¹¹At, ²²⁷Th and ²²⁵Ac based RICs (see Figure 2A, 3A, Figure 4A&B, Figure 5A and Figure 6A). We speculate this may have been due to the ascorbic acid used in the RIC preparations. The bar graphs in the corresponding figures plot the area under the curve over time and since there was no cytotoxic toxic effect of the ²¹¹At RICs, the overall effect (at low doses) was growth enhancement.

Given that the same anti-HIV antibodies and target cells were used in all experiments, it is not clear why the longer lived radiometal α-emitters were effective while the ²¹¹At-labeled mAbs were not. While there was a significant difference in calculated SAs among the ²¹¹At-mAb-B10, the ²²⁷Th-PGT126-octapa, and the ²²⁵Ac-7B2-py4pa preparations, the ratios of unlabeled to labeled antibody molecule conjugates in the preparations were similar, suggesting that a low surface ligand number was not responsible for the difference in cytotoxicity. Among all the RIC preparations tested, including those represented here (see Table 1), the unlabeled to labeled mAb ratios for ²¹¹At-mAb-B10 ranged from 57:1 to 227:1; for ²²⁷Th-PGT126-octapa it was 47:1; and for ²²⁵Ac-7B2-py4pa they ranged from 18:1 to 149:1. With these ratios it is likely that all cells had one or more α-labeled antibody attached, which should have led to at least modest specific cell killing with the ²¹¹At-7B2-B10 or ²¹¹At-PGT126-B10. One explanation for the different cytotoxic effects of the individual RICs may be the location of the RICs when they decay. Previous studies using these same mAbs as immunotoxins and antibody-drug conjugates showed them to be cytotoxic and therefore must have been internalized [25, 26]. Internalized RICs that are metabolized release their radionuclide which, in the case of halogens like ²¹¹At, may escape from the cell before decay, while radiometals, like ²²⁷Th and ²²⁵Ac, can be residualized within the cell leading to intracellular decay and more effective killing. Additionally, the subsequent decay pathways of ²²⁷Th and ²²⁵Ac yield 5 and 4 additional α-emitting daughters, respectively. Provided the daughter radionuclides remain in or near target cells, they could contribute to the superior cytotoxic effect of the ²²⁷Th- and ²²⁵Ac-based RICs.

We observed significant bystander killing of non-Env expressing cells in mixing studies where the ratio of non-target:target cells was 1:10 (Figure 6). This ratio was chosen to maximize the likelihood of detecting off-target cytotoxicity. This effect could lead to off-target organ toxicity in vivo, although the effect would likely be limited to cells directly adjacent to target cells given the short path length of α-particles. Additionally, since the HIV reservoir is composed of rare widely dispersed cells, it is unlikely that the number of cells that are vulnerable to bystander killing would be significant and lead to organ toxicity. Alternatively, limited bystander killing might be beneficial if cells adjacent to targeted cells also contain HIV proviruses due to cell-to-cell viral spread.

4.1. CONCLUSION

We have described the development of novel ²²⁷Th- and ²²⁵Ac-labeled anti-HIV α-RICs targeting the HIV envelop that are potent, specific, and cytotoxic for cells expressing surface antigens. These reagents provide the key advantages of working independently of the immune system, killing both dividing and resting cells, and they are not immunogenic, allowing for repeated rounds of treatment. Pairing these α-RICs with effective latency-reversing strategies may provide the means to deplete or eradicate the HIV reservoir.

KEY POINTS:

QUESTION:

Are anti-HIV α-radioimmunoconjugates directed against the viral envelope efficient at targeting HIV-1 infected cells?

PERTINENT FINDINGS:

anti-HIV α-radioimmunoconjugates conjugated to either ²¹¹At, ²²⁷Th- or ²²⁵Ac were tested against Env-expressing cell lines. While ²¹¹At- was associated with unspecific cytotoxicity, ²²⁷Th- and ²²⁵Ac-labeled anti-HIV α-RICs demonstrated potent and specific cytotoxicity against the target cells.

IMPLICATIONS FOR PATIENT CARE:

anti-HIV α-radioimmunoconjugates could provide a relevant therapeutic approach to eradicate target cells presenting low antigen expression, including but not limited to those infected with HIV-1.