Radiotherapy-induced Immune Response Enhanced by Selective HDAC6 Inhibition

This article is featured in Selected Articles from This Issue, p. 1363

Radiotherapy is a curative cancer treatment modality used as a single agent or in combination with surgery, chemotherapy, or immunotherapy (1). Clinical advances in radiotherapy have generally focused on improving the technical aspects of radiation delivery, beam shaping, dose fractionation, and particle therapy. Biologic responses of tumors to radiotherapy include free radical-induced DNA damage, DNA damage response cell signaling, cell-cycle arrest, and activation of the immune system (2, 3). Combining other therapies with radiotherapy and radiation-sensitizing cytotoxic drugs improves tumor control by reducing the tumor burden or clearing postsurgical microscopic residual disease. The recent successes engaging the immune system in cancer treatment with checkpoint inhibitors have supported strategies to combine radiotherapy with immune therapies as a systemic treatment strategy to improve clinical outcomes (4).

Radiation damage leads to the activation of cell death mechanisms, including cytosolic localization of DNA fragments sensed by the stimulator of interferon genes (STING) pathway and activation of adaptive immune responses against tumors (5). The concept of immune system activation after radiotherapy has been suggested by sporadic clinical observations of “abscopal effects” on distal, metastatic lesions following primary tumor irradiation, but the translation to clinical benefit has proven difficult (6). Recent successes of immune checkpoint inhibitors have boosted options for cancer treatment and generated interest in strategies that include local and systemic use of immunomodulators (7, 8). Improved survival after lung cancer treatment with immune checkpoint inhibitors and radiotherapy is a compelling stimulus to optimize the roles of radiotherapy and the immune system to enhance the effectiveness of cancer treatment (7).

Most solid malignancies exhibit a high degree of stromal and immune cell infiltration, including immunosuppressive regulatory T cells (T_reg), myeloid-derived suppressor cells (MDSC), and tumor-associated macrophages (TAM). Macrophages are critical a component of innate immune responses and key mediators of tissue homeostasis, elimination of pathogens, clearance of cellular debris, and regulation of inflammatory responses (8). Macrophages can originate from different tissues, including the yolk sac, fetal liver, and bone marrow, and once systemically disseminated, cytokines present in the tissue microenvironment (TME) determine their final phenotype (9). Macrophage polarization is a continuum that changes in response to external signals. Macrophage activation states are categorized as M1, promoting a proinflammatory, anticancer response and upregulating genes for antigen presentation and processing, or as M2, playing a critical role in wound healing, tissue regeneration, and promoting tumor progression (10). However, this is an overly simplistic classification as the phenotype can switch between these two subtypes or even adopt hybrid characteristics.

Regardless of the positive immune activation effects triggered by radiotherapy, irradiated tumors actively recruit circulating monocytes that differentiate into TAMs. TAMs, primarily M2-polarized macrophages, are strongly associated with a poor prognosis in cancer (11). TAMs secrete anti-inflammatory cytokines, induce hypoxia, have a protumor function and a low antigen cross-presentation, express immunosuppressive mediators, and exhibit detrimental effects on CD8 effector T-cell function (12). Thus, the modulation of the polarization of macrophages in irradiated tumors could be an attractive option to prevent the triggering of survival pathways activated following radiotherapy.

The emerging role of the M1/M2 ratio as a prognostic marker in cancer treatment is highlighted by recent observations that show macrophage phenotype in TME correlates with aggressiveness in most types of cancer (13). Other investigators have proposed the ratio of antitumoral M1 and protumoral M2 macrophages (M1/M2) as a potential biomarker in various malignancies (14). Importantly, recent publications have shown that the infiltration of immunosuppressive cells, including M2 macrophages, is associated with a poor prognosis after radiotherapy (15).

Histone deacetylases (HDAC), a family of 11 zinc-dependent enzymes described initially as histone modifiers, are also known to deacetylate proteins unrelated to chromatin (16). pan-HDACis (HDAC inhibitors) have thus far been used as anticancer agents in clinical settings, but critical limitations, include lower activity in solid tumors and cardiac toxicity thereby discouraging long-term therapy (17). HDACs play critical roles in key cellular functions, including cell growth, differentiation, metabolism, immune response, and sensitivity to ionizing radiation by regulating levels of protein acetylation (18, 19). Inhibitors of HDACs have roles in treating cancer, neurologic, and immune diseases (20, 21). The class IIb HDAC6 deacetylates α-tubulin, HSP-90, and interacts with the STAT3 transcription factor to regulate the immune response (22, 23). Selective HDAC6 inhibition in antigen-presenting cells, such as macrophages and dendritic cells, maintains a proinflammatory state (23). The objective of this study is to demonstrate a proof of principle that HDAC6 inhibitor–treated macrophages can be used as an adoptive cell therapy in combination with radiotherapy to improve antitumor immune responses. Here, we present a highly selective HDAC6 inhibitor, SP-2-225, for further evaluation in combination with radiotherapy as an activator of antitumor innate immunity for cancer treatment. By targeting the HDAC6 enzyme, a critical mediator of the innate immune response, a prolonged M1 macrophage polarization, and tumor growth inhibition were observed in a syngeneic immune-competent animal model.

To a solution of compound 1 (68.0 g, 343 mmol), triphenylphosphine (99.0 g, 377 mmol) and 4-hydroxy-3-methylbenzaldehyde (46.7 g, 343 mmol) in dry Tetrahydrofuran (THF) (500 mL) cooled by ice-bath, DIAD (76.3 g, 377 mmol) in THF (100 mL) was added dropwise. The resulting yellow solution was allowed to warm to room temperature and stirred for additional 20 hours (monitored by TLC). The mixture was concentrated under reduced pressure and the residue was purified by flash column on silica gel using EtOAc/hexane = 1:10 to give compound 1 as colorless solid (18.0 g, yield: 16%).

To a solution of compound 2 (18.0 g, 56.9 mmol), methyl 6-aminocaproate hydrochloride (10.3 g, 56.9 mmol) in dichloromethane (500 mL) was added sodium triacetoxy borohydride (36.2 g, 171 mmol) and was stirred at 85°C for 45 minutes. The crude material was purified by silica gel flash column to afford compound 2 (21.1 g, yield: 83%).

To a solution of compound 3 (21.1 g, 47.4 mmol) in MeOH (200 mL) and DCM (100 mL) was added DBU (7.2 g, 47.4 mmol) and NH₂OH (100 mL, 1 mol/L) in H₂O at 0°C. After addition, the reaction was stirred at 0°C for 2 hours (monitored by LC.MS). After the reaction was completed, the solid was filtered, washed by H₂O and dried to afford the desired compound SP-2-225 (12.1 g, yield: 57%).

¹H NMR (400 MHz, DMSO) δ 8.64 (s, 1H), 7.49–7.11 (m, 11H), 5.17 (s, 2H), 3.91 (s, 2H), 2.73–2.69 (m, 4H), 2.21 (s, 6H), 1.90–1.85 (m, 3H), 1.72–1.65(m, 3H).

High performance liquid chromatography (HPLC) purity: 98.0 (214 nm), 97.7% (254 nm)

Mass: m/z 447.4[M+H] ⁺

To determine plasma bioavailability of SP-2-225, three Sprague-Dawley rats were tested following intravenous injections of 10 mg/kg of SP-2-225 and 200 μL of blood were drawn via jugular venopuncture at timepoints: 0.083, 0.25, 0.5, 1, 2, 4, 8, 24 hours. Samples were stored on ice until centrifuged. Blood samples were centrifuged, and the resulting plasma and cells were separated and stored at −70°C until further analysis. Standard parameters including AUC (AUC0-t and AUC0-∞), elimination half-life (T1/2), maximum plasma concentration (Cmax) were determined. Supplementary Figure S3 summarizes the pharmacokinetics (see Supplementary Data).

The time to reach maximum plasma concentration (Tmax) and other parameters were calculated using Phoenix WinNonlin 7.0 (Pharsight). An aliquot of 20 μL blank sample was protein precipitated with 400 μL Methanol containing 100 ng/mL IS. The mixture was vortexed for 1 minute and centrifuged at 18,000 × g for 7 minutes. Transfer of 200 μL supernatant to 96-well plates was performed. An aliquot of 2 μL supernatant was injected for LC/MS-MS analysis. The T1/2 was 8.36 hours.

The SP-2-225 molecule (C28H34N2O3, 446.589 Da) was built using Avogadro computational software, and its structure was geometrically optimized using the force field MMFF94 and the steepest descent algorithm. The three-dimensional structure of the HDAC6 catalytic domain 2 (HDAC6-CD2) was obtained from the Protein Data Bank (PDB ID: 5EDU). The Zn+2 and K+1 molecules were conserved, and all other molecules were manually stripped.

The molecular docking using the protein HDAC6-CD2 (receptor) with SP-2-225 (ligand) was performed using AutoDock 4.2.6 and AutoDock tools version 1.5.6. Kollman charges were added to the receptor and ligand. The grid dimension space was limited to include the catalytic and ligand-binding amino acids in the HDAC6-CD2, including the Zn+2 molecule. Its dimension space was x-centering: 15.309, y-centering: −44.1, and z-centering: 106.533. The docking parameters were determined with the Genetic Algorithm with 100 iterations. Selection of the best docking configuration was determined by the binding energy scores and its similarity in the position with other inhibitors from the crystal structure (Trichostatin A in the 5EDU structure).

The molecular dynamics (MD) simulation was executed with Amber 16 software using the ff14SB force field with explicit solvent using the TIP3P water model (24) and a cutoff of 0.8 nm. The SP-2-225 molecular charges and amber force field parameters were obtained using antechamber and the semiempirical method AM1, and the generalized amber force field. The system was centered in a rectangular box with and neutralized with 3 Na+ and solvated in a 12 Å space between the protein mass center to the box edges. The MD simulation was done using the sander module with an energy minimization step of 1,000 steps using the steepest descent algorithm. After the energy minimization, a heating (using the isothermal-isobaric ensemble) and density equilibration MD simulation were done (500 ps each). The production of MD simulation was done for 200 ns with a time step of 2 fs with a temperature of 310 K and 1 bar of pressure using an isothermal-isobaric ensemble. The trajectory analysis was done using mdtraj (25) and pytraj (26) following the methodology proposed by Leon and colleagues (27) calculating the RMSD and RMSF, respectively. The electrostatic potential surfaces were calculated using Pymol (https://pymol.org/2/). The binding free energy between HDAC6-CD2 and SP-2-225 was calculated from additional 50 ns MD simulation after the main MD production with the MM-PBSA.py amber module.

The 4T1 triple-negative murine breast carcinoma cell line was a generous gift from Dr. Scott Abrams’ laboratory at the Roswell Park Comprehensive Cancer Center (Buffalo, NY) and SM1 murine melanoma cells were obtained from Dr. Antoni Ribas’ laboratory at the University of California, Los Angeles (Los Angeles, CA). MCF-7 breast cancer cells were obtained from ATCC. The cells were cultured in RPMI1640 medium supplemented with 10% FBS, 1% penicillin-streptomycin and incubated at 37°C with 5% CO₂. To perform HDAC inhibition and cytotoxicity assays, cells were incubated with vehicle, 0.1, 0.5, 1, 2.5, 5, 10, and 25 μmol/L concentrations of HDAC6 inhibitors Nexturastat A, SP-2-225 and pan-HDACi Panobinostat (LBH589) overnight. Assays were performed similarly on both SM1 melanoma cells and 4T1 breast cancer cells. HDAC inhibition was analyzed using the HDAC-Glo I/II Assay and Screening System (catalog no. G6420, Promega) following manufacturer's protocol. Briefly, 10,000 cells/well were seeded in a 96-well white flat clear bottom plate. After 24 hours, the plate was treated with HDACis at the indicated concentrations and incubated for 1 hour. After incubation with the HDACis, the developer was mixed with substrate and added directly to the 96-well plate. Immediately the plate was read on the SpectraMax i3x multi-mode plate reader (Molecular Devices). Cytotoxicity assays were performed with CellTox Green (Promega, catalog no. G8731) following the manufacturer's instructions, and fluorescence readings were obtained on SpectraMax i3x plate reader at excitation 485 nm and emission 520 nm wavelengths. Data analysis and plots were generated using Microsoft Excel software.

Cell lysates were prepared with RIPA buffer (Thermo Fisher Scientific, catalog no. 89900). Protease and phosphatase inhibitor (Thermo Fisher Scientific, catalog no. 78440) were added to retain posttranslation modifications on proteins in the lysate. Protein estimation was performed by bicinchoninic acid assay (Thermo Fisher Scientific, catalog no. 23225) and equal concentrations of proteins were analyzed on 4%–20% gradient SDS-PAGE (Bio-Rad, catalog no. 456-1093). Proteins were transferred from SDS-PAGE to low fluorescence polyvinylidene difluoride membranes (Bio-Rad, catalog no. 1704274) using a Trans-Blot Turbo transfer system (Bio-Rad). The membranes were blocked for 1 hour at room temperature with shaking in Odyssey blocking buffer (LI-COR, catalog no. 927-40000) followed by incubation with primary antibodies (1/1,000 dilution) at 4°C. The primary antibodies used are α-tubulin (Cell Signaling Technology, 3873), acetyl-α-tubulin (Cell Signaling Technology, 3971), and acetyl-histone H3 (Cell Signaling Technology, 9649). The membranes were washed in 1× phosphate-buffered saline with 0.05% Tween 20 (PBST) buffer (3×) followed by incubation with near-infrared fluorophore-conjugated secondary antibodies goat anti-rabbit 800 (Azure Biosystems, catalog no. AC2134) and goat anti-mouse 700 (Azure Biosystems, catalog no. AC2129) at 1/10,000 dilution for 1 hour at room temperature. The membranes were scanned on an Azure Biosystems c600 imager at near-infrared wavelengths. The images were analyzed and processed with Image Studio Lite software (version 5.2).

MCF-7 cells were plated and treated with 1 μmol/L class I HDACi SP-1-303, or 10 μmol/L HDAC6 inhibitor SP-2-225 for 24 hours. A total of 0.1% DMSO served as a negative control and 5 μmol/L proteasome inhibitor MG-132 as a positive control. Aggresomes were examined using Aggresome detection kit (Abcam, ab139486). The aggresomes were imaged using Nikon Eclipse Ts2 microscope (40× magnification). The DNA in the nucleus was counterstained with DAPI (4′,6-diamidino-2-phenylindole).

Animal studies were performed in accordance with the approval of Institutional Animal Care and Use Committee (IACUC) (Protocol# A354) at George Washington University (Washington, DC). Murine melanoma tumors were established by implanting 1 × 10⁶ SM1 cells in the right flanks of 6–8 weeks old female C57BL/6 mice. After allowing the tumors to grow to a size of approximately 200–400 mm³, the mice were randomly assigned to cohorts based on the experimental design. Initially, we tested the antitumor effect of SP-2-225 with SM1 murine melanoma model. Vehicle (PEG400/Tween80/Ethanol (70/10/20% by weight) and SP-2-225 at 25 mg/kg were administered intraperitoneal every day for 5 days per week as shown in the schematic Fig. 4A. Tumor measurements were obtained every other day to track the tumor growth kinetics until the endpoint, which is when the tumor size just exceeds 2,000 mm³.

While the implanted tumors were growing in the mice, bone marrow–derived macrophages (BMDM) were isolated for cell transplant therapy. Isolated bone marrow cells were cultured in RPMI1640 complete medium with 1% non-essential amino acids, 1% penicillin-streptomycin, and 10% FBS and incubated in 5% CO₂ at 37°C. Briefly, bone marrow from femur and tibia bones was flushed with complete RPMI media using a 3 mL syringe with 26G needle. Bone marrow was resuspended to break clumps and cultured in 10 cm culture plates with murine recombinant MCSF (20 ng/mL). On day 4, undifferentiated and floating cells were washed with PBS and replaced with fresh RPMI media. Macrophages were pretreated with 5 μmol/L of HDAC6 inhibitor SP-2-225 or vehicle prior to adding M1 polarizing factors; recombinant murine interferon gamma (Ifng; 50 ng/mL), and bacterial lipopolysaccharide (100 ng/mL) for 24 hours. The macrophages were harvested and washed thoroughly with PBS and resuspended to a cell density of 1 × 10⁶ cells in 100 μL of PBS for intratumor implantation.

To initially test the efficacy of macrophage adoptive cell therapy, mice implanted with SM1 murine melanoma tumors were randomized into the following groups: control tumors injected with PBS, tumors adoptively transferred with M1 macrophages, and tumors implanted with M1 macrophage treated with HDAC6 inhibitor SP-2-225. Macrophage adoptive cell therapy was performed weekly with intratumor implantation of either 1 × 10⁶ M1 and M1 macrophages treated with SP-2-225. The adoptive cell therapy was performed once a week till the control tumors reached the endpoint mentioned in IACUC protocol which is about 2,000 mm³. Tumor volumes are measured every other day and calculated using the formula (L × W²)/2. Once the tumor size reached the endpoint, animals were humanely euthanized.

Coculture experiments were performed with BMDMs derived from UBC-GFP [C57BL/6-Tg(UBC-GFP)30Scha/J] with constitutive expression of GFP driven by the ubiquitin promoter and BMDMs from wildtype C57BL/6 mice. GFP-expressing bone marrows cells were plated at a density of 100,000 cells per insert into Corning FluoroBlok 24-Multiwell insert system (catalog no. 351158). The insert system was then transferred into a new Falcon 24-well bottom plate (catalog no. 353047) for SP-2-225 treatments and subsequent polarization to respective phenotypes as mentioned above. The bottom plate of Corning FluoroBlok 24-Multiwell insert system was plated with 100,000 SM1 and/or Tramp-C2 cells 24 hours before the coculture, based on the experimental design and subjected to 6 Gy of gamma radiation from Cs-137 irradiator. In another iteration, 24-well bottom plates were seeded with cancer cells and 20,000 M1 or M2 wildtype BMDMs/well added 2 hours prior to subjecting to irradiation. One day after radiation, the inserts with BMDMs and 24-well bottom with cancer cells were cocultured for 24 hours. Inserts were removed and images were acquired on Molecular Devices ImageXpress Pico imager using GFP channel. GFP cell count was performed using CellReporterXpress Image aquisition and analysis software.

To test the efficacy of radiotherapy in combination with macrophage adoptive cell therapy, mice with SM1 tumors were randomized into six groups: (i) control and single-arm therapies with (ii) radiation, (iii) intratumor adoptive cell transplantation of M1 macrophages, and (iv) M1 macrophages treated with SP-2-225. Combination therapy groups were as follows: (v) radiotherapy in combination with M1 macrophage adoptive cell therapy and finally (vi) radiotherapy in combination with M1 macrophages treated with SP-2-225 adoptive cell therapy. The treatment regimen is indicated in the schematic (Fig. 6A). Cohorts marked for radiotherapy were subjected to gamma irradiation 24 hours prior to macrophage cell therapy. The reason to wait 24 hours after radiation was to enhance tumor antigens prior to intratumor macrophage implantation. The mice in the radiotherapy group were exposed to at least 12 Gy of gamma radiation with Cesium 137 irradiator by restraining the mice in a BrainTree restraint (BrainTree, catalog no. MHS1-F RF) covered with ¼” thick lead shield (catalog no. MHS1-S RF) only to expose the right flank to gamma irradiation. Additional shielding was provided with Cerrobend alloy blocks to effectively attenuate the radiation exposure to the rest of the body. The mice were monitored for tumor growth progression as mentioned above.

At the endpoint of the study, mice were humanely euthanized following the IACUC protocol. SM1 melanoma tumors were collected after necropsy. Other vital organs were inspected for differences between the treatment groups. Tumor samples were processed following the protocol as described before (reference). Briefly, the tumors were minced using a sterile surgical scalpel and digested for 45 minutes with tumor digestion buffer at 37°C with constant rotation. The number of live cells was estimated and 1 × 10⁶ cells/mL were stained with live dead Zombie Aqua assay. Fluorophore conjugated antibodies for macrophages, T cells, and natural killer (NK) cells discrimination were purchased from BioLegend unless otherwise specified. The following are antibodies used for flow cytometry to measure the activity and infiltration of NK and T cells: PerCP/Cy5.5 anti-mouse CD3(T cells; BioLegend, 100218), Alexa Flour 488 anti-mouse CD4 (CD4⁺ T cells; BioLegend, 100423), PE/Cy7 anti-mouse CD8a (CD8⁺ T cells; BioLegend, 100766), APC/Fire 750 anti-mouse CD49b (NK cells; BioLegend, 108922), Brilliant Violet 421 anti-mouse CD25 (T-cell activation; BioLegend, 102034), and Brilliant Violet 785 anti-mouse CD45.2 [T-cell activation (BioLegend, 109839)]. The flow cytometry panel to measure the activity and infiltration of myeloid cells: Brilliant Violet 421 anti-mouse/human C11b (Macrophage; BioLegend, 101236), APC anti-mouse CD80 (M1; BioLegend, 104714), PE/Cy7 anti-mouse CD206 MMR (M2; BioLegend, 141720), Brilliant Violet 630 anti-mouse CD11c (mDC; BioLegend, 117339), APC/Fire 750 anti-mouse CD45.2 (MDSC; BioLegend, 109832), PE anti-mouse CD123 [IL3 receptor (IL3R α)] (BioLegend, 106003), Brilliant Violet 603 anti-mouse Ly-6G/Ly 6C (MDSC; BioLegend, 108440), FITC anti-mouse H 2 (M2; BioLegend, 125508), and Brilliant Violet 785 anti-mouse F4/80 (Macrophage; BioLegend, 123141). After staining for 30 minutes at room temperature, cells were washed three times with 1× PBS. Samples were fixed with Life Technologies IC Fixation Buffer (FB001) from Thermo Fisher Scientific according to the manufacturer's protocol and then resuspended in FACS buffer.

The data generated in this study are available within the article and its Supplementary Data.

To synthesize SP-2-225, biphenyl benzyl alcohol (1) was combined with 4-hydroxy-3-methylbenzaldehyde via Mitsunobu coupling using DIAD. Reduced yield is attributed to the difficulty in separating unreacted biphenyl benzyl alcohol from the coupled product. However, this reaction is highly scalable to generate product 2. Compound 2 was subjected to a selective reductive amination procedure using methyl 6-aminocaproate hydrochloride and sodium triacetoxy borohydride which preferentially reduced the imide while leaving the ester intact, resulting in a high yield of compound 3 (Fig. 1A). Traditional methods of generating the hydroxamic acid functionality, such as conversion to the acid followed by coupling and subsequent deprotection of benzyl hydroxylamine or direct displacement of the ester using a large excess of aqueous hydroxylamine (30–50 eq) proved ineffective in generating the hydroxamic acid due to generation of the carboxylic acid as a degradation product and the inability to efficiently separate the carboxylic acid degradation product from the desired hydroxamic acid. A modification of a procedure using DBU and 2–3 eq of aqueous hydroxylamine cooled to 0°C for 2 hours cleanly produced the desired final compound, SP-2-225, on a multi-gram scale in relatively high yield without contamination of the carboxylic acid byproduct.

Multi-functional role of HDAC6 protein in various cellular and biological processes has attracted interest in developing targeted inhibitors that are highly specific and isoform selective. Moreover, HDAC6 is unique among the HDAC family proteins containing two catalytic deacetylase domains. However, the catalytic domain 2 (CD2) of HDAC6 confers the most deacetylase activity (22). Therefore, CD2 is the most targeted domain for developing novel HDAC6 inhibitors. Here, we represent the molecular structure of SP-2-225 (Fig. 1A) that resembles classical HDAC6 inhibitors based on a hydroxamate core (presence of a CAP, a linker, and a Zinc Binding Group - ZBG). The interaction of HDAC6-CD2 with SP-2-225 was performed in silico using Autodock and Autodock tools (Fig. 1B), and in all conformations, SP-2-225 showed a high similarity with other HDAC6 inhibitors (28) because SP-2-225 forms hydrogen bonds with histidine residues (His 610 and His 611) present in the catalytic pocket (Fig. 1D and H). In the case of His 610, Peng and colleagues showed that HDAC6 inhibitors that enhance antitumor immunity in melanoma during anti-PDL1 immunotherapy also form this interaction (29). After 200 ns MD simulation, the backbone protein structure of HDAC6-CD2 showed a stable conformation while interacting with SP-2-225 (RMSD average = 2.421 ± 0.3212 Å) shown in Supplementary Fig. S1A and RMSF by residue shown in Supplementary Fig. S1B and groups shown in Supplementary Fig. S1C. The interaction of SP-2-225 with HDAC6-CD2 was conserved during the simulation (Fig. 1E), and it showed increased interaction of SP-2-225 with the Zn⁺² molecule in the catalytic cavity inner region of the HDAC6-CD2 (Fig. 1G and I). Also, the electrostatic potential surface during the interaction of the CD2 with SP-2-225 showed the presence of negative charge areas surrounding the catalytic cavity of HDAC6-CD2 (Fig. 1C) and after the MD simulation, the HDAC6-CD2 electrostatic surface became more neutral than the initial interaction (Fig. 1F). This change in the electrostatic potential surface in the HDAC6-CD2 because of the interaction with SP-2-225 could indicate the predisposition of the protein structure to interact with the inhibitor. The free energy of interaction was calculated from an additional 50 ns MD simulation using molecular mechanics/Poisson–Boltzmann surface area (MM/PBSA) method, and the ΔG_bind was −19.4311 kcal/mol (ΔE_VDW = −31.021 kcal/mol, ΔE_ele = 69.669 kcal/mol, ΔE_PB = −54.591 kcal/mol, and ΔE_POLAR = −3.4883 kcal/mol). The calculated free energy of interaction between SP-2-225 with HDAC6-CD2 is in agreement with the free energy of interaction of other HDAC6 inhibitors such as Nexturastat (−12.421 kcal/mol), Rocilinostat (−18.909 kcal/mol), Tubacin (−25.987 kcal/mol), and Tubastatin (−9.802 kcal/mol; ref. 28). Structures of SP-2-93, SP-1-161, SP-2-213, and SP-2-225 are shown in Supplementary Fig. S1D. Overall, based on the MD simulations and modeling, the data thus far indicated that SP-2-225 is a superior HDAC6 inhibitor.

HDAC inhibitory properties of SP-2-225 were assessed in two different ways; we first assessed using whole cell extract containing all HDAC enzymes to determine pan-HDAC inhibition. Second, analyses of individually purified HDAC isoforms. The IC₅₀ values of SP-2-225 were determined by direct enzymatic inhibition assays in two class I isoforms, HDAC1 and HDAC3, as well as class IIa HDAC6. In both the pan-HDAC and class I isoform inhibition assays, SP-2-225 yielded IC₅₀ values in the micromolar range, while the HDAC6 IC₅₀ was in the low nanomolar range (67 nmol/L). In contrast, the well-known pan-HDACi suberoylanilide hydroxamic acid (SAHA) produced IC₅₀ values for all three HDAC isoforms in the low nanomolar range. These data demonstrated substantial selectivity of SP-2-225 for HDAC6 over the class I HDAC isoforms, with the ratio of pan-HDAC inhibition to HDAC6 inhibition over 250-fold compared with that of SAHA, which is 3.6-fold. (Fig. 2A). Therefore, SP-2-225 is a highly selective HDAC6 inhibitor rather than a non-specific, potent HDAC6 inhibitor.

As inhibitory concentrations determined from purified enzymatic assays can be significantly lower than those eventually observed in cell-based assays, substrate acetylation was determined in MCF-7 breast cancer cells for isoform selectivity. Acetylation of the HDAC6 substrate α-tubulin, and histone H3, the substrate for HDAC1 and HDAC3, were assessed by Western blot analysis. The acetylation profile generated by treatment of cells with 10 μmol/L SP-2-225 were compared with those produced by treatment with two pan-HDACis, SAHA and SP-1-161 at 1 μmol/L. A lower concentration of 1 μmol/L for pan-HDACis was used to avoid cytotoxicity. As seen in Fig. 2B, both pan-HDACis increased the acetylation of α-tubulin and histone H3 equally or demonstrated increased acetylation of histone H3 lysine 9 (Ac-H3K9). In contrast, SP-2-225 significantly increased the acetylation level of α-tubulin but not histone of H3, corroborating the specificity seen with purified HDAC proteins.

We also determined that the deacetylase activity of SP-2-225 was similar to the well-characterized HDAC6i Nexturastat A (30, 31) in SM1 murine melanoma cells (Fig. 2C). In addition, the cytotoxic effect of SP-2-225 in vitro was slightly lower than Nexturastat A (Fig. 2D), indicating an optimal profile to perform in vitro studies at 5 μmol/L without inducing cell death. These studies were also reproduced using murine 4T1 breast cancer cells (Supplementary Fig. S2A and S2B), suggesting the potential use of this compound in other types of solid tumors. We used the pan-HDACi LBH589 as a positive control for deacetylase activity. Moreover, LBH589 has been reported to induce a direct cytotoxic effect even at nanomolar concentrations (31).

To evaluate whether SP-2-225 could influence the functional phenotype of macrophages, we polarized BMDMs to the M2 phenotype in the presence of SP-2-225. As expected, after 24 hours of treatment with the polarizing cytokines IL4 and IL13, these macrophages highly expressed the M2 marker Arginase 1 (Arg1). However, the expression of this marker was significantly reduced in the presence of SP-2-225 (Fig. 2E), indicating that this drug could effectively reduce the polarization of protumoral M2 macrophages. Similarly, the expression of anti-inflammatory cytokine Il10 and negative regulator of STAT3 signaling Socs3, which are expressed by the M1 phenotype were suppressed by SP-2-225 (Supplementary Fig. S4A and S4B), indicating that SP-2-225 could effectively sustain the antitumoral properties of M1 macrophages.

To advance SP-2-225 into preclinical studies in animal models, we determined its pharmacokinetics in Sprague-Dawley rats treated with intravenous injections of 10 mg/kg. The results of consecutive sample collections up to 24 hours indicated a good pharmacokinetic profile with a maximum plasma concentration (Cmax) of 3,605 ng/mL and elimination half-life (T1/2) of 8.365 hours (Fig. 2F; Supplementary Fig. S3). The long half-life of the compound suggests that it is metabolically stable and would be an excellent candidate for every other day or every third day treatment to maintain an effective level of the compound in the bloodstream. Overall, the data provide a strong rationale for in vivo testing of SP-2-225 for immunomodulatory antitumor properties.

HDAC6 plays a critical function in the clearance of aggresomes by activating the proteasome protein degradation system (32). To reinforce SP-2-225’s capacity to inhibit HDAC6, we exposed MCF7 breast cancer cells to SP-2-225 and the proteasome inhibitor MG-132 as a positive control and stained cells for the presence of aggresomes. As expected, direct inhibition of proteasomes with MG-132 caused significant aggresome formation after 24 hours of treatment (Fig. 2G). The HDAC6 inhibitor SP-2-225 also caused aggresomal formation. Further analysis of SM1 murine melanoma cells with increasing concentrations of SP-1-161 (pan-HDACi) and SP-2-225 (HDAC6 inhibitor) and analogs of SP-2-225 (SP-2-93, SP-2-213, SP-2-223) overnight was performed. Treatment with SP-2-225 caused an increase in only acetyl-tubulin but did not affect the histone H3 acetylation status, whereas SP-1-161, which is a pan-HDACi, resulted in increased acetylation status of both histone H3 and tubulin (Fig. 2H). These data further functionally validate that SP-2-225 is a highly specific inhibitor that targets only HDAC6.

Previous work from our group demonstrated that selective HDAC6is reduce tumor growth in syngeneic murine tumor models with an intact immune system (31, 33). A similar outcome of tumor reduction was observed with the syngeneic SM1 murine melanoma model after 26 days of treatment with SP-2-225 at a dose of 25 mg/kg administered intraperitoneally as shown in the schematic (Fig. 3A). Compared with the vehicle-treated group, treatment with SP-2-225 significantly decreased the tumor burden, as evidenced by the tumor growth curves in Fig. 3B–D. Immunophenotyping of tumor-infiltrated immune cells by flow cytometry did not indicate a significant increase in the infiltration of either CD8⁺ or CD4⁺ T cells (Fig. 3E and F). However, the treatment with SP-2-225 increased the infiltration of CD8⁺ effector memory T cells (CD8 EM), suggesting activation of adaptive immune responses. Further evaluation of TAMs in the TME revealed a significant reduction of the M2 phenotype (Fig. 3G), which led to a sharp increase in the M1/M2 macrophage ratio, indicating that targeting HDAC6 reduced the anti-inflammatory and protumoral TME. Consistent with Fig. 2E, SP-2-225 also demonstrated significant immunomodulatory effects by decreasing M2 macrophages in vivo, thereby resulting in an increased M1/M2 ratio.

Autologous transplantation and adoptive transfer of macrophages differentiated from blood monocytes have been tested in clinical trials to treat autoimmune diseases, transplant rejection, cardiac diseases, and cancers. However, this approach has provided only modest therapeutic benefits in patients (34). These clinical trials were initiated before fully understanding the complexity of macrophage phenotypes, their ability to alter their phenotypes and polarization in response to various cytokines and the tumor/tissue environment cues, and the extent of survival of the transplanted macrophages. It is now known that the TME effectively reprograms macrophages after transplantation. For example, antitumor M1 macrophages are reprogrammed toward the protumoral M2 phenotype after transplantation due to the dominant anti-inflammatory niche promoted by tumor cells in the TME (34).

We previously reported that treatment with HDAC6is reduced the polarization of macrophages toward the protumor M2 phenotype and increased M1/M2 ratio. We performed an adoptive cell therapy with macrophages in SM1 syngeneic melanoma model as shown in the schematic (Fig. 4A). Remarkably, the treatment of M1 macrophages with SP-2-225 before adoptive transfer in SM1 melanoma-bearing mice significantly reduced tumor growth (Fig. 4B–E), suggesting a potent effect of this HDAC6i on the polarization and function of macrophages. As indicated by the tumor growth curves, adoptive cell therapy with SP-2-225–treated macrophages suppressed melanoma tumor growth compared with M1 and vehicle control cohorts. Effective polarization of BMDMs to M1 macrophages was validated for increased expression of proinflammatory cytokine IL1B by qRT-PCR (Supplementary Fig. S4C). The intraperitoneal administration and intraadoptive cell therapy data are interpreted to show that SP-2-225 is an effective modulator of antitumor immune responses by affecting macrophage phenotype.

TAMs originate from tissue-resident macrophages and predominantly from tumor-infiltrated circulating monocytes that subsequently polarize to TAMs ((35). Moreover, longitudinal analyses of metabolic and proteomic profiles of plasma from patients with prostate cancer with radiotherapy indicated that patients experiencing disease progression had elevated levels of M2 metabolites (36). Therefore, we investigated the effect of SP-2-225 on the recruitment of macrophages toward irradiated cancer cells. Using an in vitro coculture system, we demonstrated that GFP-expressing M1 or M2 BMDMs from UBC-GFP mice in the transwell insert migrated toward irradiated cancer cells present in the bottom of a 24-well plate. As shown in Fig. 5A, more M2-polarized macrophages migrated toward non-irradiated cancer cells compared with the M1-polarized GFP macrophages suggesting that cancer cells inherently secrete soluble factors that induce tumor-promoting M2 macrophage migration. However, pretreatment with SP-2-225 significantly reduced the number of M2 macrophages migrating toward SM1 melanoma cells. With irradiated SM1 cells (6 Gy) in Fig. 5B, we observed a slight increase in the number of migrated M2 macrophages, and SP-2-225 significantly decreased the migration of M2 macrophages. The plate templates are shown in Supplementary Fig. S5A and S5B. We further validated the effect of SP-2-225 mediated suppression of M2 macrophage migration toward cancer cells using Tramp-C2 prostate cancer cells (Fig. 5C and D), suggesting that HDA6 inhibition in M2 macrophages reduced migration toward cancer cells.

To further mimic the TME consisting of tumor-resident macrophages and cancer cells, we cocultured either M1 or M2 BMDMs from wildtype C57BL/6 mice along with Tramp-C2 prostate cancer cells in a 24-well plate. The plates were either exposed to 6 Gy radiation or not followed by coculture with GFP-expressing M1 or M2 BMDMs in transwell inserts, as shown in the schematic in Fig. 5E. GFP-expressing M1 macrophages exhibited increased migration toward irradiated cancer cells than non-irradiated cancer cells when cocultured with wildtype M1 macrophages. Furthermore, SP-2-225 treatment of GFP M1 macrophages did not affect migration (Fig. 5F). However, Tramp-C2 prostate cancer cells cocultured with M2 macrophages, increased migration of GFP M1 macrophages regardless of irradiation and SP-2-225 did not influence M1 macrophage migration (Fig. 5G). The results indicate that SP-2-225 treatment of M1 macrophages does not affect M1 macrophages migration toward cancer cells. The plate templates are shown in Supplementary Fig. S6A and S6B. On the other hand, M2 GFP macrophages from the inserts migrated at a higher number than M1 GFP macrophages toward Tramp-C2 cells cocultured with either wildtype M1 or M2 macrophages. SP-2-225 treatment of M2 GFP macrophages in the insert significantly decreased migration toward non-irradiated cancer cells, but the reduction was nonsignificant with irradiated cancer cells (Fig. 5H). With GFP M2 macrophages in the insert, coculture with Tramp-C2 cancer cells, and wildtype M2 macrophages in the bottom well, there was increased migration of GFP M2 macrophages that was scuttled with SP-2-225 treatment (Fig. 5I). The plate templates are shown in Supplementary Fig. S6C and S6D. Irradiation slightly increased the migration of GFP M2 macrophages toward Tramp-C2 cancer cells cocultured with compared with non-irradiated cancer cells. However, the numbers were significantly higher than irradiated cancer cells alone (Fig. 6D compared with Fig. 6I), suggesting that TAMs present in the TME significantly affect infiltration of macrophages into the tumor. The data thus far support the notion that irradiated tumors attract immunosuppressive M2 macrophages, and the HDAC6 inhibitor, SP-2-225 can inhibit this process.

Radiotherapy has been used in various solid malignancies, combined with primary, adjuvant, and salvage treatment strategies (37). Radiotherapy was traditionally thought to induce cell death through DNA damage and apoptosis. It is now accepted that radiotherapy also causes immunogenic death of tumor cells that generate neoantigens (38), leading to an immune-stimulatory effect and subsequent remission of tumors even outside the radiation field by the abscopal effect (39). Several recent publications have demonstrated that anti-inflammatory protumoral M2 macrophages are associated with poor prognosis in solid tumors (14, 40). Notably, the infiltration of M2 macrophages after radiotherapy has also been associated with bad prognosis and tumor relapse (15, 41).

Considering the strong effect of SP-2-225 on tumor growth and phenotypic composition of the TME, we hypothesized that the treatment of irradiated tumors with HDAC6i reprogrammed M1 macrophages could diminish the protumoral TME encountered in the aftermath of radiotherapy. This hypothesis was tested using the SM1 murine melanoma syngeneic model, where tumors were irradiated (12 Gy) and subjected to macrophage adoptive transfer 24 hours after radiotherapy (Fig. 6A). As shown in Fig. 6B, irradiated tumors were significantly smaller than non-irradiated tumors indicating that radiotherapy was effective. The intratumoral implantation of M1 macrophages untreated or pretreated with SP-2-225 after irradiation resulted in further reduction of tumor growth compared with respective non-irradiated tumors, indicating that M1 macrophages improve antitumor immunity. Notably, the tumor reduction was associated with a high M1/M2 macrophage ratio (Fig. 6C) and improved infiltration of CD8⁺ T cells (Fig. 6D). RT+SP-2-225–treated M1 macrophage therapy also increased CD8 effector and central memory cells. (Fig. 6E and F). Compared with Fig. 3D where SP-2-225 systemic therapy did not affect CD8⁺ T-cell infiltration, radiotherapy combined with SP-2-225–treated M1 macrophage adoptive cell therapy significantly increased the CD8⁺ T-cell response. Consistent with other in vivo studies, we did not see a significant effect on CD4⁺ T cells (Fig. 6E, H, and I). Taken together, these data support the hypothesis that HDAC6is pretreated macrophages could improve the beneficial effect encountered when transplanting antitumor M1 macrophages in irradiated tumors.

Early publications have reported the potential role of the host immune system after radiotherapy (42). However, the participation of specific immune populations and their positive or negative impact on the curative effect of radiotherapy is under investigation. Furthermore, the abscopal effect of reducing tumor growth outside the radiation field enhanced interest in understanding the role of radiotherapy in activating antitumor immunity (43). However, a systematic review of cases of abscopal effects indicated the phenomenon to be very uncommon (44), suggesting that host-mediated mechanisms could be the main protagonists in activating the immune system after radiotherapy. Therefore, one may conclude that complementing radiotherapy with other treatment modalities, such as immunotherapy, could potentially boost antitumor immunity (1). There is currently an increased interest in developing novel combination therapies with radiotherapy toward this approach. Our work demonstrates a proof of principle for personalized medicine cancer treatment by combining radiotherapy with immunotherapy in a preclinical model tumor system. Physical targeting of tumors was offered by focal radiotherapy delivery, and molecular targeting was achieved in conjunction with the activation of the innate immune system using the novel HDAC6 inhibitor SP-2-225. The highly selective HDAC6 inhibitor has minimal effect on the class I HDAC family of enzymes, as demonstrated by enzymatic and cell-based assays. In fact, MCF7 cells exposed to 10 μmol/L SP-2-225 caused large increases in tubulin acetylation but no increases in histone acetylation. The inhibition of HDAC6 by SP-2-225 was further validated through the phenotypic identification of aggresome formation. While SP-2-225 failed to substantively inhibit the growth of murine breast or melanoma cells in vitro, the growth of syngeneic SM1 tumors was significantly reduced through treatment with SP-2-225. Because of the lack of cytotoxicity coupled with the ability of SP-2-225 to dramatically reduce tumor growth in murine models, we believe the mechanism of growth inhibition to be indirect, potentially functioning through the animal's immune system. Therefore, using class selective HDACis is a superior approach to pan-HDACis due to the inherent cytotoxicity of pan-HDACis. Because of relatively low toxicity, HDAC6 inhibitors can be used in long-term anticancer therapy. Moreover, HDAC6 inhibitors exert antitumor effects by immunomodulation of TME. Furthermore, HDACis at large (45) and particularly HDAC6 inhibition with tubacin in bladder cancer has demonstrated radiosensitization effect (46).

Macrophages are an important component of TME, and they function as tumor-promoting (M2) or tumor suppressing (M1) factors depending on the inflammatory status of the tumor (47). Soluble factors such as cytokines and chemokines mobilize circulating monocytes to the TME. Radiotherapy is known to induce the release of CCL2, a monocyte chemoattractant (48) in addition to CSF-1, which is essential for monocyte to macrophage differentiation (49). We demonstrated with in vitro coculture experiments that SM1 melanoma and Tramp-C2 prostate cancer cells recruit macrophages which could be through soluble factors, and this recruitment was predominantly tumor promoting M2-polarized macrophages. We were able to demonstrate that HDAC6 inhibition with SP-2-225 significantly reduced the migration of M2 macrophages toward irradiated cancer cells.

The M1/M2 ratio has been correlated to patient's prognosis, where a lower M1/M2 ratio is correlated to an unfavorable outcome, and a higher M1/M2 ratio is associated with a favorable outcome (50). However, one should notice that such a correlation is usually context and cancer dependent. Therefore, altering the phenotype of TAMs is of current interest. In our report, using the SM1 murine melanoma model with an active immune system, systemic treatment through intraperitoneal administration of SP-2-225 significantly increased M1/M2 ratio and was associated with a reduced tumor burden compared with the vehicle-treated group. Furthermore, adoptive cell therapy with SP-2-225–treated macrophages was similarly effective in reducing the tumor burden with a very high translational value. In the subsequent combination therapy with radiotherapy, HDAC6 inhibitor–treated M1 macrophages not only retained the M1 phenotype, which is evident by a sharp increase in the M1/M2 ratio but also increased the infiltration of CD8⁺ effector T cells. Surprisingly, there was also an increase in the number of CD8⁺ effector memory and central memory cells, suggesting that radiotherapy may have improved the neoantigen repertoire of tumor cells, leading to effective antigen presentation through transplanted M1 macrophages. However, further investigation of antigen-mediated immune responses with HDAC6i-treated M1 macrophages is warranted. Overall, we demonstrated the effectiveness of a novel combination therapy with a high translational potential where radiotherapy combined with HDAC6i-treated M1 macrophage adoptive cell therapy effectively develops lasting antitumor immunity against solid tumors.

S. Grindrod reports a patent for US 2020/0071288 A1 issued. N. Aghdam reports a patent for Hdac6-activated macrophages, compositions, and uses thereof WO2020264437A1 pending and licensed. M. Jung reports other support from Shuttle Pharmaceuticals, Inc outside the submitted work. A. Dritschilo reports other support from Shuttle Pharmaceuticals, Inc. during the conduct of the study; other support from Shuttle Pharmaceuticals, Inc. outside the submitted work; in addition, A. Dritschilo has a patent for U.S. Patent No: 11,407,723 Selective HDAC inhibitors for the treatment of human disease issued to No license involved. Patents are owned by Shuttle Pharmaceuticals, Inc.. No disclosures were reported by the other authors.

S.K.R. Noonepalle: Conceptualization, data curation, software, formal analysis, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. S. Grindrod: Conceptualization, data curation, software, formal analysis, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. N. Aghdam: Conceptualization, software, formal analysis, supervision, investigation, project administration. X. Li: Data curation, software, formal analysis, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing. M. Gracia-Hernandez: Data curation, software, formal analysis, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing. C. Zevallos-Delgado: Data curation, software, formal analysis, investigation, visualization, methodology, writing–original draft, writing–review and editing. M. Jung: Conceptualization, resources, data curation, software, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. A. Villagra: Conceptualization, resources, data curation, supervision, funding acquisition, writing–original draft, project administration, writing–review and editing. A. Dritschilo: Conceptualization, resources, supervision, funding acquisition, writing–original draft, project administration, writing–review and editing.

Agency: NIH

Identifying Number: 1R01CA249248-01A1

Agency: Cancer Research Institute

Identifying Number: 228514

SP-2-225 and analogs were provided under MTA by Shuttle Pharmaceuticals, Inc

Powered@SouthernGPU: This research was partially supported by the supercomputing infrastructure of the Southern GPU Cluster - Fondequip EQM150134. Studies in this article were performed with the support of George Washington University's core facilities including Animal Research Facility, Flow cytometry Core and gamma irradiator.