Combining Metformin and Drug-Loaded Kidney-Targeting Micelles for Polycystic Kidney Disease

Abstract

Introduction

Autosomal dominant polycystic kidney disease (ADPKD) is the most common inherited kidney disease that leads to eventual renal failure. Metformin (MET), an AMP-activated protein kinase (AMPK) activator already approved for type 2 diabetes, is currently investigated for ADPKD treatment. However, despite high tolerability, MET showed varying therapeutic efficacy in preclinical ADPKD studies. Thus, newer strategies have combined MET with other ADPKD small molecule drug candidates, thereby targeting multiple ADPKD-associated signaling pathways to enhance therapeutic outcomes through potential drug synergy. Unfortunately, the off-target side effects caused by these additional drug candidates pose a major hurdle. To address this, our group has previously developed kidney-targeting peptide amphiphile micelles (KMs), which displayed significant kidney accumulation in vivo, for delivering drugs to the site of the disease.

Methods

To mitigate the adverse effects of ADPKD drugs and evaluate their therapeutic potential in combination with MET, herein, we loaded KMs with ADPKD drug candidates including salsalate, octreotide, bardoxolone methyl, rapamycin, tolvaptan, and pioglitazone, and tested their in vitro therapeutic efficacy when combined with free MET. Specifically, after determining the 40% inhibitory concentration for each drug (IC₄₀), the size, morphology, and surface charge of drug-loaded KMs were characterized. Next, drug-loaded KMs were applied in combination with MET to treat renal proximal tubule cells derived from Pkd1flox/-:TSLargeT mice in 2D proliferation and 3D cyst model.

Results

MET combined with all drug-loaded KMs demonstrated significantly enhanced efficacy as compared to free drugs in inhibiting cell proliferation and cyst growth. Notably, synergistic effects were found for MET and KMs loaded with either salsalate or rapamycin as determined by Bliss synergy scores.

Conclusion

Together, we show drug synergy using drug-loaded nanoparticles and free MET for the first time and present a novel nanomedicine-based combinatorial therapeutic approach for ADPKD with enhanced efficacy.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12195-022-00753-9.

Introduction

Autosomal dominant polycystic kidney disease (ADPKD) is the most prevalent genetic kidney disorder that affects 12.5 million patients globally.^8,27 ADPKD is caused by mutations in the PKD1 or PKD2 genes and is characterized by progressive, fluid-filled cyst formation leading to kidney enlargement and eventual kidney failure.^7,31,43,45 Currently, tolvaptan is the only ADPKD drug approved by the FDA to slow down renal cystogenesis.⁹ However, tolvaptan has been reported to induce off-target side effects including polyuria and liver toxicity.^24,41,60 In addition to tolvaptan, other drugs such as metformin (MET), an FDA-approved drug for type 2 diabetes, are currently under consideration for ADPKD treatment.²¹ MET is in phase 3 clinical trials for ADPKD³⁷ and is known to inhibit mitochondrial respiratory chain complex 1, thus activating AMP-activated protein kinase (AMPK)⁵⁰ which leads to inhibition of mammalian target of rapamycin (mTOR), a kinase promoting cellular proliferation and cyst growth (Fig. 1).⁵⁰ Although MET was shown to be well tolerated in ADPKD patients,^6,48,51 the therapeutic potency of MET was modest and varied in ADPKD animal studies.^26,40,53 Thus, recent efforts have been put forth into combining MET with other ADPKD drug candidates to enhance therapeutic effects.^26,29,64

Figure 1.

Schematic of signaling pathways regulated by MET and six drugs tested herein [salsalate (SAL), rapamycin (RAP), tolvaptan (TOL), octreotide (OCT), bardoxolone methyl (BAR), and pioglitazone (PIO)]. All six drugs target a different signaling pathway or regulate the same pathway through distinct mechanisms compared to MET.

Herein, we selected a library of ADPKD drug candidates, salsalate (SAL), rapamycin (RAP), tolvaptan (TOL), octreotide (OCT), bardoxolone methyl (BAR), and pioglitazone (PIO), to be combined with MET; all six drugs are either currently in clinical trials for ADPKD or have shown promising preclinical results.^21,26 As shown in Fig. 1, all drug candidates either target a different signaling pathway or regulate the same pathway with distinct mechanisms compared to MET. In addition to MET, which is an indirect AMPK activator, SAL is a prodrug dimer of salicylate which directly activates AMPK through interaction with the binding domain of the AMPK β1 isoform.^18,50 RAP directly inhibits mTOR signaling by blocking the mTOR kinase domain.²⁵ Apart from AMPK/mTOR pathway, other ADPKD candidates aim to target cyclic AMP (cAMP) signaling, which is known to stimulate cellular proliferation and cyst formation in ADPKD.¹⁶ For instance, TOL, the FDA-approved ADPKD drug, antagonizes vasopressin V2 receptor, downregulating intracellular cAMP.³ OCT also targets cAMP signaling by binding to somatostatin receptors which inhibit the generation of cAMP.²¹ Additionally, BAR is an activator of the nuclear factor erythroid 2-related factor (Nrf2) pathway, which has been shown to attenuate cyst formation by modulating inflammation in ADPKD.⁵² Finally, PIO is an agonist for the peroxisome proliferator activated receptor (PPAR-γ), which reduces cystogenesis via inhibiting expression of the cystic fibrosis transmembrane conductance regulator (CFTR) and extracellular signal-related kinase (ERK).^4,38

Despite some success in vivo, these six drugs can cause off-target adverse effects in patients.^2,62 Similar to TOL, treatments of SAL, BAR, RAP, and PIO were all reported to induce liver injury.^{22,24,28,36,57} BAR was also found to cause muscle spasms,⁴² while PIO imposes a risk of hypoglycemia, for which the long-term safety is still being investigated.^5,34 Additionally, gastrointestinal symptoms (e.g., diarrhea, nausea, or vomiting) were observed in patients receiving SAL, OCT, and BAR.^15,19,42 Chronic treatment of OCT was also reported to inhibit patients’ gallbladder emptying, thereby leading to cholelithiasis.⁴⁴

To mitigate the challenge of off-target effects, our group has developed a kidney-targeted drug delivery system based on peptide amphiphile micelles (KMs) by incorporating the megalin-binding peptide ((Lys-Lys-Glu-Glu-Glu)₃-Lys) ((KKEEE)₃K) onto the micelle surface.^20,59,63 Upon intravenous injection into wildtype mice, we found KMs displayed significant kidney accumulation and high biocompatibility.^20,59 We verified that KMs bind to megalin, a multi-ligand receptor expressed in renal proximal tubule epithelial cells,^12,20,59 which form cysts in early ADPKD.¹⁷ Importantly, due to the amphiphilic structure of micelles, hydrophobic drugs, including our selected six candidates, can be efficiently loaded into the lipid core of KMs via hydrophobic interaction.³² Hence, drug-loaded KMs can potentially be used as promising candidates in combination with MET to slow down renal cyst growth.

In the present study, we first investigated the dose response for all six drugs in free drug form (SAL, OCT, BAR, RAP, TOL, and PIO) combined with MET on mouse ADPKD cell proliferation. As the proliferation inhibition rates of all drugs reach 40% at the maximum dose, we selected the 40% inhibitory concentration (IC₄₀) as the dose for each drug to be incorporated into KMs. Next, we applied combinations of drug-loaded KMs and MET on mouse ADPKD cells in 2D proliferation and 3D cyst model to evaluate its in vitro therapeutic effects. Overall, we present a combinatorial therapeutic approach for ADPKD using drug-loaded nanoparticles and MET, which provides a novel prospect of renal drug delivery with potentially enhanced therapeutic efficacy and drug synergy.

Materials and Methods

Cell Culture

The polycystic renal proximal tubular epithelial cell lines (PKD1 null) were clonally derived from Pkd1flox/-:TSLargeT mice following in vitro Cre recombinase transfection. PKD1 null cells were maintained and passaged in DMEM/F12 media (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 2% FBS, 1 × ITSG, and ~ 2 nM of tri-ido-sodium salt at 37 °C and 5% CO2. The media was changed every 2–3 days.

Dose Determination of Free Drugs

MET, BAR, and TOL were purchased from Sigma-Aldrich (USA); SAL and OCT were purchased from MedChemExpress (USA); RAP was purchased from Santa Cruz Biotech (USA); PIO was purchased from BioVision Inc. (USA). PKD1 Null cells were trypsinized and resuspended in 200 μL media/well to achieve approximately 3000 cells/well and were seeded in a 96-well plate. The cells were allowed to grow for 24 h to reach ~ 30% confluence and were treated with different doses of free drug combinations for 48 h. Cell proliferation was assessed with a 3-(4,5- dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) cell proliferation colorimetric assay (Biovision, Milpitas, CA, USA) (n ≥ 3). The MTS reagent was added at 48 h post-treatment followed by absorbance measured at 490 nm using a Varioskan LUX plate reader (Thermo Fisher Scientific, Waltham, MA, USA). Cell proliferation was calculated using the following equation

$$\text{Cell}\text{proliferation}\left(\text{\%}\text{PBS}\right)=\left(\frac{{\text{Absorbance}}_{\text{sample}}}{{\text{Absorbance}}_{\text{PBS}}}\right)\times 100$$ 1

The 40% inhibitory concentration (IC₄₀) was determined through the nonlinear regression analysis using GraphPad Prism 8 software (GraphPad Software, San Diego, CA).

Synthesis of Drug-Loaded KMs

The kidney-targeting (KKEEE)₃K peptide was synthesized using standard Fmoc-mediated solid phase peptide synthesis on an automatic PS3 peptide synthesizer (Protein Technologies, Tucson, AZ, USA) with rink Amide resin (Protein Technologies, Tucson, AZ, USA). A cysteine was added to the N-terminus of the peptide sequence for the thioether linkage reaction. The peptides were then cleaved from the resin with 94:2.5:2.5:1 volume ratio of trifluoroacetic acid:1,2-ethanedithiol:H₂O:triisopropylsilane. Cleaved peptides were precipitated and washed several times with ice cold diethyl ether, dissolved in Milli-Q water, lyophilized, and stored at − 20 °C. The crude peptides were purified by reverse-phase high performance liquid chromatography (HPLC, Shimadzu, Kyoto, Japan) on a C18 column (Phenomenex, Torrance, CA, USA) at 55 °C with 0.1% formic acid in acetonitrile/water mixture. The purified peptides were characterized using matrix-assisted laser desorption ionization time-of-flight mass spectral analysis (MALDI-TOF, Fig. S1a) (Autoflex speed, Bruker, Billerica, MA, USA). Then, the purified cysteine-containing peptides were conjugated to 1,2 distearoyl-sn-glycero-3-phosphoethanolamineN-[maleimide(polyethylene glycol)-2000], or DSPE-PEG(2000)-maleimide (Avanti Polar Lipids, Alabaster, AL, USA) via a thioether linkage by mixing an equimolar amount of the lipid and pure peptide in Milli-Q water (pH 7.2) at room temperature for over 48 h with gentle agitation. The resulting DSPE-PEG(2000)-(KKEEE)₃K was purified by HPLC on a C4 column (Phenomenex, Torrance, CA, USA) as described above and characterized using electrospray ionization mass spectrometer (ESI MS, Fig. S1b) (Shimadzu, Kyoto, Japan).

DSPE-PEG(2000) amphiphiles and drugs were self-assembled into drug-loaded micelles via thin film evaporation. The appropriate DSPE-PEG(2000) amphiphiles and drugs were dissolved in methanol or chloroform and evaporated with nitrogen gas to form thin films. The resulting thin films were dried overnight under vacuum and hydrated at 80 °C for 30 min. Micelles prepared for characterization were hydrated in Milli-Q water, whereas micelles prepared for in vitro study were hydrated in phosphate buffered saline (PBS). The micelle solutions were vortexed and sonicated as needed to obtain a clear solution and allowed to cool to room temperature. All KMs were composed of an amphiphile molar ration of 50:50 consisting of DSPE-PEG(2000)-(KKEEE)₃K:DSPE-PEG(2000)-methoxy and were loaded with determined doses of drugs.

Characterization of Drug-Loaded KMs

Loading Efficiency

The standard curves of six drugs were first obtained at the wavelengths of maximum absorbance (Fig. S2, SAL at 310 nm, OCT at 277 nm, BAR at 253 nm, RAP at 278 nm, TOL at 274 nm, and PIO at 270 nm) via a NanoDrop One microvolume UV–VIS spectrophotometer (Thermofisher Scientific, Waltham, MA, USA). The drugs were dissolved completely in dimethyl sulfoxide (DMSO). while the absorbance was assessed. The drug-loaded KMs (100 μM micelle concentration) were prepared in Milli-Q water, and the unincorporated free drug was filtered via a 0.22 µm PES membrane filter (Thermofisher Scientific, Waltham, MA, USA). The absorbance of the encapsulated drugs was measured at the wavelengths of maximum absorbance via NanoDrop. The concentration of the encapsulated drug was calculated using the standard curve, and the loading efficiency was obtained using the following formula (n = 4).

$$\text{Loading}\text{efficiency}(\%)=\frac{\text{Weight}\text{of}\text{drugs}\text{loaded}\text{in}\text{micelle}\left(\text{mg}\right)}{\text{Weight}\text{of}\text{total}\text{drugs}\left(\text{mg}\right)}\times 100$$ 2

Drug Release Profile

After micelle formation, free drug was separated via a 0.22 µm PES membrane filter. Measurements of free drug release from KMs were made at select timepoints (30 min, 1, 2, 6, 12 h). Then, KMs were disassembled using a 10:1 volume ratio of DMSO, releasing the encapsulated drug into solution. After vortex and sonication, absorbance of the solution was measured via Nanodrop at the wavelength of absorbance of each drug. The drug concentration was calculated from absorbance using a standard curve for each drug (n = 3).

Dynamic Light Scattering (DLS) and Zeta Potential

100 µM of micelle solution was prepared in Milli-Q water and the unincorporated free drug was filtered. The diameter and zeta potential of the micelle were measured using Zetasizer Ultra (Malvern Instruments, Malvern, UK) immediately after micelles were hydrated from thin films. All measurements were carried out at 25 °C after equilibrating for 5 min (n = 4).

Transmission Electron Microscopy (TEM)

100 µM of micelle solution was prepared in Milli-Q water and the unincorporated free drug was filtered. The micelle solution was placed on a 200-mesh carbon TEM grid (Ted Pella, Redding, CA) for 5 min. Excess liquid was wicked away, and the grid was washed with Milli-Q water before being negatively stained with 2 wt% uranyl acetate solution (Polysciences, Warrington, PA). After 2 min, the stinging solution was wicked away, and the grid was washed with Milli-Q water. The grid was kept in the dark under room temperature and imaged on a JEOL JEM-2100F TEM (JEOL, Ltd., Tokyo, Japan).

In Vitro Evaluation of PKD1 Null Cell Proliferation

The clonal renal proximal tubular epithelial cells generated from Pkd1flox/-:TSLargeT mice (PKD1 null cells) were cultured in a 96-well plate using the same method as described above. Cells were treated with 100 µM of empty KM, 1 mM MET, the IC₄₀ of each drug (SAL, OCT, BAR, RAP, TOL, or PIO) as determined above, 1 mM MET combined with the IC₄₀ of each drug as determined above, drug-loaded 100 µM KM, or 1 mM MET combined with drug-loaded 100 µM KM for 48 h. Cell proliferation was assessed with MTS assays using the same method as in the Dose Determination of Free Drugs section (n ≥ 4).

The Bliss synergy score was calculated using the following equation

$${\text{S}}_{\text{bliss}}=\left(y_c-\left(y_a+y_b-y_a{y}_b\right)\right)\times 100$$ 3

where $y_a$ and $y_b$ are the observed inhibition rates with treatment alone and treatment b alone, and $y_c$ is the observed inhibition rates with treatment a and b combined. The inhibition rates were calculated as (1 − cell proliferation %).

In Vitro Evaluation of PKD1 Null Cyst Growth

50 μL of Matrigel (Corning Life Sciences Inc., NY, USA) was added to each well in a 96 well plate and solidified in a 37 °C incubator for 15 min. PKD1 null cells were trypsinized and resuspended with 150 μL of 2% Matrigel™ in assay medium to achieve approximately 3000 cells/well (Day 0), and the cells were grown for 2 days. Then, cells were treated with 100 µM of empty KM, 1 mM MET, the IC₄₀ of each drug (SAL, OCT, BAR, RAP, TOL, or PIO) as determined above, 1 mM MET combined with the IC₄₀ of each drug as determined above, drug-loaded 100 µM KM, or 1 mM MET combined with drug-loaded 100 µM KM on Day 3, 6, 9, and 12. On each treatment day, medium was displaced, and cells in each well were refed with 150 μL assay medium supplemented with 2% Matrigel™. On Day 8, 11, and 14, cysts were imaged by a microscope (Leica DMi8, Leica, Wetzlar, Germany), and cyst diameter was measured by ImageJ (n ≥ 20). The cyst volume was calculated as $\frac{4}{3}\pi r^3$, where r is diameter/2.

Statistical Analysis

Data are expressed as means ± SEM and analyzed by a Student’s t-test to compare means of pairs. Analysis of variance (ANOVA) was used to evaluate significant differences among three or more means. Statistical analyses were performed using GraphPad Prism 8 software. p ≤ 0.05 was considered to be statistically significant.

Results

Dose Determination of Free Drugs

In this study, we aimed to evaluate the therapeutic effect of MET combined with drug-loaded KMs using renal proximal tubular epithelial cell lines generated from Pkd1flox/-:TSLargeT mice (PKD1 null cells) as an in vitro ADPKD model. Before synthesizing drug-loaded KMs, we evaluated the free drug dose response for all six drugs combined with MET on kidney cell proliferation, in order to select the optimal dose for nanoparticle incorporation.

Before testing the six drugs, the effects of MET on PKD1 null cell proliferation were investigated. Specifically, PKD1 null cells were incubated with different concentrations (0.3–2.5 mM) of MET for 48 h (Fig. 2a). These concentrations were chosen based on a previous in vitro study where a range of 0.6–5 mM of MET was applied to inhibit proliferation of human polycystic kidney cells.⁶⁴ As the dose increased from 0.6 to 1 mM, cell proliferation significantly decreased to ~ 85% of the PBS level (p ≤ 0.05, Fig. 2a). Beyond 1 mM of MET, the inhibitory effect on cell proliferation did not change significantly as shown in Fig. 2a. Therefore, 1 mM MET was selected for subsequent drug combination studies.

Figure 2.

Dose response of drug combinations on proliferation of the renal proximal tubule cells derived from Pkd1flox/-:TSLargeT mice (PKD1 null). (a) Dose response of MET (0.3–2.5 mM) on PKD1 null cells upon incubation for 48 h. Dose response of 1 mM MET combined with (b) SAL (0.1–0.8 mM), (c) OCT (0.02–0.2 mM), (d) BAR (0.1–0.4 µM), (e) RAP (0.005–0.1 µM), (f) TOL (0.01–0.1 mM), and (g) PIO (0.01–0.1 mM) upon incubation for 48 h. Line: IC₄₀ (proliferation = 60% normalized to PBS treatment) in (b)–(g). *p ≤ 0.05, **p ≤ 0.01.

Next, to evaluate the combined effects of the drug candidates and MET on cell proliferation, we incubated PKD1 null cells with each of the six drugs combined with 1 mM MET. The dose range for each drug was chosen based on previous studies.^{23,26,35,46,47,49} As found in Figs. 2b–2g, cell proliferation gradually declined when treated with an increasing drug dose combined with 1 mM MET. At the maximum dose chosen for each drug, the proliferation rate was 58.3% or less of PBS group when combined with 1 mM MET (Figs. 2b–2g, proliferation rate ranges from BAR: 11.1 ± 5.8% to PIO: 58.3 ± 7.6% of PBS group). The 40% inhibitory concentrations (IC₄₀, proliferation = 60% of PBS group) for all six drugs combined with 1 mM MET were estimated and found to be 487.7 μM SAL, 63.53 μM OCT, 140.3 nM BAR, 4.60 nM RAP, 71.29 μM TOL, and 91.20 μM PIO (Figs. 2b–2g). The IC₄₀ of each drug was chosen to be loaded into KMs for the subsequent studies.

Synthesis and Characterization of Drug-Loaded KMs

The kidney-targeting peptide, (KKEEE)₃K, which binds to megalin expressed on proximal tubule cells which form cysts during early ADPKD,⁶³ was modified with a cysteine on the N-terminus and conjugated to DSPE-PEG(2000)-maleimide via a thioether linkage to construct peptide amphiphile molecules (Fig. 3a). Micelles consisting of 50:50 molar ratio of DSPE-PEG(2000)-(KKEEE)₃K:DSPE-PEG(2000)-methoxy were co-assembled with each drug by thin film hydration, resulting in drugs encapsulated within the hydrophobic core (Fig. 3a).

Figure 3.

Schematic of the synthesis of drug-loaded KM and representative transmission electron microscopy images. (a) (KKEEE)₃K peptide was conjugated to DSPE-PEG(2000) via a thioether linkage, and the amphiphile monomers and drugs were co-assembled to form drug-loaded KMs. TEM images of (b) KMs without drug loading, KMs loaded with (c) SAL, (d) OCT, (e) BAR, (f) RAP, (g) TOL, and (h) PIO show spherical morphology.

As shown in Table 1, the loading efficiencies for all drugs were higher than 50% in KMs at 100 μM micelle concentration. KMs without drug loading had an average diameter of 11.1 nm as measured by DLS, whereas the diameter was slightly increased for drug-loaded KMs (11.1 ± 0.4–15.0 ± 0.7 nm, Table 1). The zeta potentials of all KMs were negative (Table 1), which was expected due to the incorporation of the anionic (KKEEE)₃K peptide.^20,59 All KMs were also characterized via TEM which showed a monodisperse population of spherical nanoparticles (Figs. 3b–3h). There is no change in spherical morphology upon the addition of all drugs into KMs. Finally, the drug release profile showed cumulative release ranging from 69.5 ± 0.2% (SAL) to 98.0 ± 0.2% (OCT) by 12 h (Fig. S3), confirming the release of free drugs from KM.

Table 1.

Loading efficiency, size, and zeta potential of micelles.

Micelle	Loading efficiency (%)	Diameter (nm)	Zeta potential (mV)
KM	–	11.1 ± 0.2	− 4.95 ± 0.7
KM_SAL*	73.1 ± 0.4	13.5 ± 0.9	− 3.4 ± 0.6
KM_OCT*	52.3 ± 0.7	12.0 ± 0.2	− 3.3 ± 0.4
KM_BAR*	75.4 ± 3.1	11.4 ± 0.5	− 13.3 ± 1.3
KM_RAP*	92.4 ± 0.3	11.1 ± 0.4	− 10.6 ± 0.5
KM_TOL*	63.6 ± 0.4	12.1 ± 0.4	− 27.1 ± 1.5
KM_PIO*	67.2 ± 1.1	15.0 ± 0.7	− 33.2 ± 0.4

Data are expressed as means ± SEM

KM kidney-targeting peptide amphiphile micelle; SAL salsalate; OCT octreotide; BAR bardoxolone methyl; RAP rapamycin; TOL tolvaptan; PIO pioglitazone

*Drug-loaded KM

Effects of MET + Drug-Loaded KMs on PKD1 Null Cell Proliferation

After synthesizing and characterizing drug-loaded KMs (KM_drugs), we combined KM_drugs with 1 mM MET and evaluated cell proliferation of PKD1 null cells after 48 h post-treatment (Fig. 4a).

Figure 4.

Effects of MET + KM_drug on PKD1 null cell proliferation. (a) Schematic depicting how drug-loaded KMs are combined with MET and incubated with PKD1 null cells cultured in 96-well plates for this study. Cells were treated with PBS, empty KMs, MET alone, free drugs alone, MET + free drugs, KM_drugs alone, and MET + KM_drugs for 48 h, where drugs are (b) SAL, (c) OCT, (d) BAR, (e) RAP, (f) TOL, and (g) PIO. Cell proliferation was detrmined using a MTS assay, normalized to PBS treatment. *p ≤ 0.05, **p ≤ 0.01, ****p ≤ 0.0001.

Effects of MET + KM_Drugs vs KM_Drugs on Cell Proliferation

First, to evaluate the effects of MET in combinations, we compared the cell proliferation under combination treatments (MET + KM_drugs) and treatments of KM_drugs alone. As shown in Figs. 4b and 4e respectively, cell proliferation was reduced by 15.9 ± 5.6% for MET + KM_SAL relative to KM_SAL (P ≤ 0.05) and was reduced by 18.0 ± 3.6% for MET + KM_RAP relative to KM_RAP (p ≤ 0.01), demonstrating significantly enhanced inhibitory effects by MET + KM_SAL/RAP relative to KM_SAL/RAP alone. However, there was no significant difference in cell proliferation between groups treated with MET + KM_drug and KM_drug for the other four drugs (OCT, BAR, TOL, and PIO, Figs. 4c, 4d, 4f, and 4g).

Effects of MET + KM_drugs vs MET + Free Drugs on Cell Proliferation

Next, we compared the inhibitory effects on cell proliferation when MET is combined with KM_drugs or free drugs. As shown in Figs. 4b–4g, treatments of MET + KM_drugs displayed significantly enhanced inhibitory effects on cell proliferation relative to their MET + free drug combination counterparts for all six drug candidates. The enhancement of inhibition rate for MET + KM_drugs relative to MET + free drugs ranges from 8.5 ± 2.4% (BAR, p ≤ 0.05, Fig. 4d) to 31.0 ± 3.6% (RAP, p ≤ 0.0001, Fig. 4e). Without being combined with MET, the cell proliferation was also significantly reduced for KM_drug treatments as compared to their free drug counterparts (Figs. 4b–4g), further demonstrating the enhanced inhibitory effects for all six drugs upon encapsulation into nanoparticles. PKD1 null cells treated with empty KMs showed no effects on proliferation and were similar to the PBS control (Figs. 4b–4g).

Synergies of Drug Combinations in Inhibiting Cell Proliferation

To compare and analyze the synergistic effects of free drug combinations as well as combinations of MET and KM_drugs, we calculated the Bliss synergy score for all combinations using previously reported methods (Table 2).^30,33 The Bliss synergy score indicates the percentage of the observed inhibition rate of proliferation upon treatment with A + B drug combination, and synergism is represented as above (positive) the predicted inhibition rate of proliferation by drug A and B alone while antagonism is below (negative).³⁰ The Bliss synergy score for MET + SAL was 9.0 ± 1.2 (Table 2), corresponding to a 9.0 ± 1.2% of inhibition rate beyond the predicted effect.³⁰ The score for MET + KM_SAL was 8.7 ± 1.3 (Table 2), indicating that the synergistic effect found between free MET and SAL was not compromised upon SAL incorporation in KM. Interestingly, the synergy score for MET + RAP was marginally negative at − 1.5 ± 0.8 (Table 2), suggesting a slight antagonism of the combination.³⁰ However, when MET was combined with KM_RAP, the score shifted up to 10.0 ± 0.9 (Table 2, p < 0.0001), showing a significant increase in synergistic effect for MET combined with RAP upon encapsulation into nanoparticles. The synergy scores were positive for MET combined with free OCT (2.1 ± 1.1), TOL (3.5 ± 0.5) and PIO (8.2 ± 1.0), but were negative for MET combined with KM_OCT (− 1.4 ± 0.7), KM_TOL (− 4.2 ± 0.5), and KM_PIO (− 4.8 ± 0.5), indicating a reduction in synergy (change from synergism to antagonism) for MET combined with OCT, TOL, or PIO upon encapsulation into nanoparticles (Table 2). Finally, the combinations of both MET + BAR (− 4.4 ± 1.3) and MET + KM_BAR (− 7.0 ± 0.6) were shown to be antagonistic (Table 2).

Table 2.

Bliss synergy scores of drug combinations.

Drug	MET + free drug	MET + KM_drug*
SAL	9.0 ± 1.2	8.7 ± 1.3
OCT	2.1 ± 1.1	− 1.4 ± 0.7
BAR	− 4.4 ± 1.3	− 7.0 ± 0.6
RAP	− 1.5 ± 0.8	10.0 ± 0.9
TOL	3.5 ± 0.5	− 4.2 ± 0.5
PIO	8.2 ± 1.0	− 4.8 ± 0.5

The Bliss synergy score indicates the percentage of the observed inhibition rate of proliferation upon treatment with A + B drug combination, and synergism is represented as above (positive) the predicted inhibition rate of proliferation by drug A and B alone while antagonism is below (negative).³⁰

Data are expressed as means ± SEM

MET metformin; KM kidney-targeting peptide amphiphile micelle; SAL salsalate; OCT octreotide; BAR bardoxolone methyl; RAP rapamycin; TOL tolvaptan; PIO pioglitazone

*Drug-loaded KM

Effects of MET + Drug-Loaded KMs on PKD1 Null Cyst Growth

Next, to assess the therapeutic potential of MET + KM_drug in reducing renal cyst growth, we applied the same treatments to a 3D cyst model. PKD1 null cells were suspended in the Matrigel matrix (day 0) and treated on day 3, 6, 9, and 12. PKD1 null cells start to form spherical cystic structures by day 5–6.⁶¹ As shown in Figs. 5 and 6, all treatments of six drugs significantly reduced the cyst size compared to the PBS control.

Figure 5.

Representative brightfield images of PKD1 null cysts treated with PBS, empty KMs, MET alone, free drugs alone, MET + free drugs, KM_drug alone, and MET + KM_drug for all six drug candidates. PKD1 null cells were placed in the Matrigel on day 0 to form cysts, and treatments were applied on day 3, 6. Brightfield images were captured on day 8. Scale bar: 30 µm.

Figure 6.

Effects of MET + KM_drug on PKD1 null cyst diameter measured on day 8. Cysts were treated with PBS, empty KMs, MET alone, free drugs alone, MET + free drugs, KM_drug alone, and MET + KM_drug, where drugs are (b) SAL, (c) OCT, (d) BAR, (e) RAP, (f) TOL, and (g) PIO. *p ≤ 0.05, **p ≤ 0.01, *** p ≤ 0.001, ****p ≤ 0.0001.

Effects of MET + KM_drugs vs KM_drugs on Cyst Growth

To evaluate the effects of MET in combinations on reducing cyst growth, we compared the cyst diameters measured on day 8 under combination treatments (MET + KM_drugs) and treatments of KM_drugs alone (Figs. 5, 6). As shown in Figs. 6a and 6d, cyst diameters were reduced by 2.6 ± 0.6 µm for MET + KM_SAL relative to KM_SAL (p ≤ 0.0001, Fig. 6a) and were reduced by 1.3 ± 0.5 µm for MET + KM_RAP relative to KM_RAP (p ≤ 0.05, Fig. 6d), demonstrating a greater reduction in cyst growth by MET + KM_SAL/RAP compared to KM_SAL/RAP alone. There was no significant difference found between MET + KM_drug and KM_drug for the other four drugs (OCT, BAR, TOL, and PIO, Figs. 6b, 6c, 6e, 6f), which was similar to their effects on cell proliferation (Figs. 4c, 4d, 4f, 4g).

Effects of MET + KM_drugs vs. MET + free Drugs on Cyst Growth

Next, we compared the effects on reducing cyst growth when MET is combined with KM_drugs or free drugs. The cyst diameters were measured on day 8, 11, and 14 and showed consistently smaller diameters for cysts treated with MET + KM_drug or KM_drugs than those treated with the free drug counterparts for all six drug candidates (Figs. 6, S4a–f), which was similar to their effects on cell proliferation (Figs. 4b–4g). We also calculated the change in cyst volumes from day 8 to day 14 and found significantly less enlargement for cysts treated with MET + KM_drugs than those treated with MET + free drugs (Fig. S4g–l). Overall, these results demonstrate a significantly enhanced effect on reducing cyst growth for all six drugs upon encapsulation into KMs. No difference in cyst size was found between empty KMs and PBS treatment (Figs. 6, S4a–f).

Discussion

In recent years, many small molecule drugs have been proposed for ADPKD therapy,^21,26 but off-target effects and ineffective delivery of drug candidates remain an issue.^2,62 Previously, peptide amphiphile micelles have emerged as a promising nanomedicine platform for various applications, including drug delivery for ADPKD.^{10,11,20,55,59} Our previous work on kidney-targeting peptide amphiphile micelles, or KMs, confirmed renal accumulation in vivo.^20,56,59 Herein, we explored the therapeutic potency of drug-loaded KM combined with MET for ADPKD. Specifically, we incorporated IC₄₀ of six drug candidates (OCT, SAL, BAR, RAP, TOL, and PIO) into KMs and applied them in combination with MET on PKD1 null cells cultured 2D and as a 3D cyst model and quantified the effects on cell proliferation and cyst growth, respectively.

For all six drugs in combination with MET, we observed significantly enhanced therapeutic effects for drug-loaded KMs relative to their free drug counterparts in both in vitro models (Figs. 4b–4g, 6, S4), which may be partly attributed to the renal-targeting and megalin-binding ability of KMs.^59,63 Megalin was also reported to play a central role in the endocytic process of its ligands,^12,54,58 and future studies will characterize and confirm whether KMs can enhance internalization and intracellular trafficking thereby enabling increased drug uptake.

Among six combinations of MET and KM_drugs, we found synergies for MET and KMs loaded with SAL or RAP in inhibiting cell proliferation (Table 2). MET + KM_SAL and MET + KM_RAP also showed significantly enhanced effects in reducing cyst growth as compared to KM_SAL and KM_RAP alone (Figs. 6a, 6d, S4g, j). As mentioned, MET is an indirect AMPK activator by inhibiting mitochondrial function, while SAL can directly activate AMPK.^18,50 Activated AMPK inhibits mTOR, whereas RAP is a direct inhibitor of mTOR signaling.^25,50 Therefore, both SAL and RAP regulate the same pathway (AMPK and mTOR, respectively) as MET, but with distinct mechanisms (Fig. 1). Previous works have demonstrated synergies between MET and SAL in inhibiting cancer cells and liver lipogenesis due to their distinct mechanisms for both activating AMPK^14,39; this may also explain the observed synergistic effect in PKD1 null cells when MET is combined with SAL or RAP upon encapsulation into nanoparticles (Table 2). Future studies will evaluate the effects of combining MET, KM_SAL, and KM_RAP on cyst growth to evaluate additional benefits on the AMPK/mTOR pathway.

The synergy scores for MET combined with OCT, TOL, and PIO were positive (synergism) at free drugs but decreased to negative (antagonism) upon incorporation into KMs (Table 2). While improved therapeutic effects were observed for MET + free OCT, TOL, or PIO relative to free drugs alone, no significant difference was found between MET + KM_drug and KM_drug for these three drugs (Figs. 4c, 4f, 4g, 6b, 6e, 6f). Different from SAL and RAP which target the same pathways as MET, OCT (cAMP), TOL (cAMP), and PIO (PPAR-γ) each regulates a distinct pathway from MET (Fig. 1).^3,4,21 Upon encapsulating into nanoparticles, the therapeutic effects of OCT, TOL, and PIO were significantly enhanced, making drug-loaded KMs play predominate roles in combinations and potentially block the effects by MET, which may lead to the observed antagonism (Table 2). Hence, KMs loaded with OCT, TOL, and PIO would have enhanced therapeutic effects but may not be ideal candidates combined with MET.

Notably, BAR was the only candidate, as either free drug or KM_drug, that exhibited antagonism with MET (Table 2). MET + BAR or MET + KM_BAR didn’t show differences in inhibiting cell proliferation and reducing cyst size as compared to BAR or KM_BAR alone (Figs. 4d, 6c). BAR is a known activator of Nrf2, which is a key regulator in maintaining mitochondrial function,¹³ while MET is known to activate AMPK by inhibiting mitochondrial function.⁵⁰ The mutual interference of mitochondrial function by two drugs may lead to the observed antagonistic therapeutic effect (Table 2). Therefore, both free BAR and KM_BAR may not be ideal candidates to be combined with MET.

Despite enthusiasm, some aspects of the system in this study should be further optimized. For example, MET was applied as a free drug in the current study, but can be incorporated as part of KMs in future studies. Although MET is hydrophilic and thus cannot be loaded into the lipid core via hydrophobic interaction, the amine group of MET can be utilized to conjugate MET as part of the hydrophilic corona on micelle surface.¹ Taken together, we investigated the therapeutic potential of the combinatorial therapies of MET and kidney-targeting micelles loaded with six different drug candidates for ADPKD. We demonstrated the significantly enhanced in vitro therapeutic efficacy of drug-loaded KMs in comparison to free drugs. Furthermore, we noticed the potential synergistic therapeutic effect exhibited by combinations of MET and KM loaded with SAL and RAP, both of which regulate the same signaling pathway as MET but with distinct mechanisms. Future studies assessing this combinational therapy in vivo will more fully elucidate its potential benefits in enhancing therapeutic efficacy and limiting off-target side effects.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

KM

Kidney-targeting peptide amphiphile micelle

MET

Metformin

SAL

Salsalate

OCT

Octreotide

BAR

Bardoxolone methyl

RAP

Rapamycin

TOL

Tolvaptan

PIO

Pioglitazone

KM_drug

Drug-loaded kidney-targeting peptide amphiphile micelle