Visual Overview
Abstract
Lysosomal sequestration may affect the pharmacokinetics, efficacy, and safety of new basic lipophilic drug candidates potentially impacting their intracellular concentrations and tissue distribution. It may also be involved in drug-drug interactions, drug resistance, and phospholipidosis. However, currently there are no assays to evaluate the lysosomotropic behavior of compounds in a setting fully meeting the needs of drug discovery. We have, therefore, integrated a set of methods to reliably rank order, quantify, and calculate the extent of lysosomal sequestration in rat hepatocytes. An indirect fluorescence-based assay monitors the displacement of the fluorescence probe LysoTracker Red by test compounds. Using a lysosomal-specific evaluation algorithm allows one to generate IC50 values at lower than previously reported concentrations. The concentration range directly agrees with the concentration dependency of the lysosomal drug content itself directly quantified by liquid chromatography–tandem mass spectrometry and thus permits a quantitative link between the indirect and the direct trapping assay. Furthermore, we have determined the full pH profile and corresponding volume fractions of the endo-/lysosomal system in plated rat hepatocytes, enabling a more accurate in silico prediction of the extent of lysosomal trapping based only on pKa values as input, allowing early predictions even prior to chemical synthesis. The concentration dependency—i.e., the saturability of the trapping—can then be determined by the IC50 values generated in vitro. Thereby, a more quantitative assessment of the susceptibility of basic lipophilic compounds for lysosomal trapping is possible.
Introduction
Many drugs show lipophilic properties in combination with moderate-to-high basicity. This property combination is often associated with pH-driven lysosomal sequestration, which can have a strong impact on the distribution, efficacy, and safety of such drugs since it may lead to high concentrations in lysosome-rich tissues such as lung, liver, kidney, and spleen (MacIntyre and Cutler, 1988a; Ndolo et al., 2012). To undergo lysosomal sequestration a compound needs to 1) be membrane permeable, and 2) have a basic moiety that is ionizable at acidic pH. Indeed, the major driving force into the lysosome is the pH gradient between the neutral cytosol (pH 7.0–7.2) (Berezhkovskiy, 2011; Hallifax and Houston, 2012; Poulin et al., 2012) and the acidic lysosomal matrix (pH 4.5–5) (Feng and Forgac, 1992; Kornhuber et al., 2010; Mindell, 2012). While being able to enter the lysosome via passive diffusion, the acidic environment within the lysosome causes a protonation of basic groups. This ionization converts the formerly neutral molecule into a charged molecule, which significantly reduces its permeability across the lipid bilayer, resulting in trapping of the cationic form (Fig. 1). Besides the lysosome, any cellular compartment with a lower pH than the cytosol (e.g., early and late endosomes) is in principle eliciting this trapping mechanism. For simplification, all acidic cellular compartments are hereinafter referred to as lysosomes. Despite the typical minor volume of <2% of the cellular volume (with exceptions in certain cell types, e.g., macrophages of up to 8%) (Ufuk et al., 2015), the contribution to cellular drug uptake can be significant since trapping can theoretically lead to as much as 160,000-fold higher drug concentrations within the lysosome relative to the cytosol (MacIntyre and Cutler, 1988a,b). Besides contributing to the distribution into lysosome-rich tissues, lysosomal sequestration also affects the intracellular localization of drugs. This may be beneficial if the therapeutic target is inside the lysosome, but in most cases the drug is actually drawn away if the site of action is in the cytosol or cell nucleus, thereby influencing its efficacy. In addition, excessive lysosomal accumulation may affect drug safety by impairing the normal physiologic function of lysosomes. Indeed, this is one of the main mechanisms of drug-induced phospholipidosis (Reasor et al., 2006). Furthermore, lysosomal trapping has also been implicated in drug-drug interactions. Due to the limited trapping capacity of lysosomes, two coadministered lysosomotropic drugs may compete for lysosomal accumulation, which in turn could lead to elevated cytosolic concentrations and hence may elicit adverse drug effects; however, thus far, this was seen only in preclinical studies and at rather high doses (Daniel and Wójcikowski, 1999b; Funk and Krise, 2012). Finally, lysosomal trapping has also been associated with cancer drug resistance. For instance, sunitinib can cause augmented lysosome biogenesis in certain cancer cells. This in turn may lead to increased lysosomal sequestration, which then may result in inadequate target exposure after continuous treatment with lysosomotropic anticancer drugs. This has been reported as a mechanism of cancer resistance in several approved drugs and is also implicated to result in cross-resistance of lysosomotropic agents (Gotink et al., 2015; Zhitomirsky and Assaraf, 2016). Looking at the potential impact of these implications, early information on lysosomal trapping in drug discovery will increase our understanding of drug distribution both at the level of body tissues and subcellular disposition, thereby supporting the selection of new drug candidates with improved efficacy and safety profile in relation to their intracellular distribution.
In this paper, we present a set of methods that were established to assess the lysosomotropic properties of basic compounds in drug discovery with particular suitability for lead optimization and candidate selection and profiling. We have developed two complementary experimental methods: 1) an indirect fluorescence-based method that utilizes the fluorescent probe LysoTracker Red DND-99 (LTR) to screen and rank order compounds for lysosomal trapping, and 2) a direct quantification method using liquid chromatography–tandem mass spectrometry (LC-MS/MS) to quantify the intralysosomal versus cellular drug content. We have furthermore derived a quantitative link between both methods. Together with in-depth characterization of the endo-/lysosomal system in cultivated rat hepatocytes, we were able to accurately predict the hepatocellular drug distribution to lysosomes.
Materials and Methods
Chemicals and Reagents.
Ammoniochloride, chloroquine, diclofenac, fluoxetine, monensin sodium, and imipramine were purchased from Sigma-Aldrich (St. Louis, MO) and propranolol was purchased from Research Biochemical International (Natick, MA). Olaparib was obtained from Active Biochem Ltd. (Hong Kong, China). LTR, Lysosensor Yellow/Blue DND-160, and Hoechst 33342 were purchased from Life Technologies (Carlsbad, CA). Sodium chloride, potassium chloride, HEPES, EGTA, and calcium chloride were purchased from Sigma-Aldrich. Acetonitrile and methanol were purchased from Honeywell Specialty Chemicals Seelze GmbH (Seelze, Germany).
Isolation of Rat Hepatocytes.
Fresh hepatocytes were isolated from male Han:Wistar rats (Janvier Laboratories, Le Genest-Saint-Isle, France). Animals were anesthetized via intraperitoneal injection of xylazine/ketamine (1:1). The liver was perfused in situ with 200 ml buffer 1 (4.2% NaCl, 0.3% KCl, 1.2% HEPES, and 0.19% EGTA in bi-distilled water, pH 7.2) and subsequently with 180 ml buffer 2 (3.9% NaCl, 0.5% KCl, 0.7% CaCl2, and 2.4% HEPES in bi-distilled water, pH 7.2) containing 13,500 U of collagenase type II (Worthington Biochemical Corporation, Lakewood, NJ) via the portal vein. After perfusion the liver was excised and the tissue was disassociated by scraping with forceps. The hepatocytes were transferred into hepatocyte medium [20 ml glutamine (Gibco, Grand Island, NY), 500 µl insulin, and 20 µl glucagon in 500 ml William’s Medium E (Sigma-Aldrich, St. Louis, MO)] and purified via Percoll (Sigma-Aldrich) centrifugation according to Kreamer et al. (1986) with hepatocyte medium as the washing solution. In the final step the purified hepatocytes were diluted in hepatocyte medium containing 5% fetal calf serum (Sigma-Aldrich).
Indirect Fluorescence-Based Assay for Lysosomotropism and Cytotoxicity Assessment.
Drug stock solutions for the assay were prepared in dimethylsulfoxide (Sigma-Aldrich), from which the loading solutions were diluted with William’s Medium E in a concentration range of 0.5–100 µM for each drug. LTR (to stain the lysosomes) and Hoechst 33342 (to assess the cytotoxic drug effects) were added to each loading solution in final concentrations of 50 nM and 4 µg/ml, respectively. Negative control contained no drug and positive control contained 25 mM NH4Cl to elevate lysosomal pH, abolishing the pH gradient and thereby the accumulation of LTR. As the cell system, freshly isolated rat hepatocytes were seeded (200,000 cells/well) in clear 24-well Biocoat Collagen I plates (Corning Incorporated, Corning, NY) and allowed to adhere for 24 hours at 37°C, 5% CO2. Each well was washed twice with prewarmed phosphate-buffered saline (PBS) (Gibco) and cells were subsequently incubated with 300 µl loading solution for 40 minutes at 37°C, 5% CO2 in order to accommodate both rapidly and slowly permeating compounds. (It should be noted that incubation time may have to be adjusted for other cell types.) The loading solution was removed and replaced with 500 µl of fresh William’s Medium E. Fluorescence was then immediately captured with an Axiovert 200 microscope (Zeiss, Jena, Germany) equipped with an AxioCam MRm. Image acquisition was performed at six random locations within each well in the rhodamine and DAPI channels. The microscope settings used are specified in Table 1. The lysosomotropic character of each drug was determined over the whole concentration range in duplicates. The assay was replicated twice in a reduced form for concentrations relevant for lysosomal trapping, resulting in n = 6 measurements.
Data Analysis and Evaluation.
Fluorescence images were analyzed with Axiovision 4.8.2 (Zeiss, Jena, Germany). The cell viability was evaluated by counting cell nuclei in the DAPI channel, with a decreasing number of nuclei indicating cytotoxic drug effects. The assessment of lysosomal drug sequestration by LTR fluorescence reduction was analyzed in the rhodamine channel. In fluorescence images each pixel has a possible intensity from 0 to 4096. The brightness of lysosomes was manually identified and two thresholds were applied. (It should be noted that thresholds need to be adjusted individually for different setups.) The entire lysosomal system was analyzed by measuring any fluorescence intensity ≥500, discarding the background/cytosol, and serving as a surrogate for the cell number in the images. For lysosomal peak intensities the second threshold of ≥ 1000 was used. Figure 2 visualizes LTR intensities with applied thresholds of a control cell and a schematic visualization of applies thresholds. Due to the variability of the number of cells per image, the proportion (plyso) of lysosomal peak intensities () in the lysosomal system () was calculated for every image. To detect alterations in the lysosomal LTR fluorescence, it was normalized to the plyso of the respective negative controls ():(1)The results of all measurements were pooled and a logistic regression on plyso/norm against compound concentration was performed in Origin 2018 (OriginLab Corporation, Northampton, MA) or Sigmaplot 13 (Systat Software, Inc San Jose, CA) to generate the IC50 values.
Direct Quantification of Lysosomal Drug Content.
To identify the relation between LTR displacement and lysosomal drug accumulation, propranolol and imipramine accumulation was measured 1) in control hepatocytes and 2) hepatocytes with inactive lysosomes as well as 3) in hepatocytes additionally treated with LTR to evaluate any interaction on drug accumulation. Each compound was tested three times over a concentration range of 1–100 µM in hepatocytes from three different rats 1) containing the test compound only, additionally containing either 2) 25 µM Monensin or 3) 50 nM LTR. Freshly isolated rat hepatocytes were cultured and plated as described previously. The plates were washed twice with prewarmed PBS and cells were incubated with 300 µl loading solution (t0) for 40 minutes at 37°C, 5% CO2. After incubation, the overlaying medium (t40) was sampled. The cells were washed twice with ice-cold PBS. Methanol containing internal standard (0.4 µM) was added to each well and the sealed plate was shaken for 3 minutes at 750 rpm to lyse the cells. The lysate was transferred to a 96-well analytical plate, in which t0 and t40 were diluted in methanol containing internal standard. The analytical plate was shaken for 15 minutes at 750 rpm and subsequently centrifuged for 15 minutes at 3700 rpm and 4°C for protein precipitation. Samples were then analyzed by LC-MS/MS.
To determine the number of cells in the assay, an additional 24-well plate was seeded and cultured as described previously. The plate was washed twice with prewarmed PBS and hepatocyte medium containing 4 µg/ml Hoechst 33342 was added to the wells. The fluorescence was measured after 10-minute incubation with an ImageXpress Micro (Molecular Devices, San Jose, CA) using a 10× objective and a DAPI filter cube. In each well 16 images were acquired. Cell nuclei were counted using ImageJ (version 1.43u; Wayne Rasband, National Institutes of Health, US).
Analytical Method.
Samples were analyzed by liquid chromatography-tandem mass spectrometry using an Agilent 1290 Infinity System (Agilent Technologies, Santa Clara, CA) comprising a G4220A binary pump, G1316C column compartment, and G7167B multisampler linked to an AB Sciex API4000 mass spectrometer with electrospray ionization (AB Sciex Germany GmbH, Darmstadt, Germany). All compounds were detected in positive multiple reactions monitoring mode with N-(4-chlorophenyl)-2-[(4-pyridinylmethyl)amino]-benzamide (Sigma-Aldrich) as an internal standard. An Ascentis Express C18 column (30 × 2.1 mm, 2.7 µm particle size; Sigma-Aldrich) was used with mobile phases (A) water with 0.1% acetic acid and (B) acetonitrile with 0.1% acetic acid and a linear rising gradient from 10% to 95% B over 0.6 minutes with a flow of 1 ml/min, followed by a washback with a flow of 1.2 ml/min to 10% B over 0.2 and 0.4 minutes of equilibration with a flow of 1 ml/min. Multiple reactions monitoring transitions for propranolol and imipramine were 260.3/116.2 and 281.3/86.1 with elution times of approximately 0.37 and 0.42 minutes, respectively. The internal standard eluted after 0.48 minutes with a multiple reactions monitoring transition of 337.8/210.9.
Data Evaluation.
Liquid chromatography mass spectrometry quantification results were corrected for recovery and the counted number of cells in the assay under the assumption of a similar lysosomotropic character of metabolites and a linear relation between cell number and accumulated drug. The accumulated amount of drug in lysosomes (Alyso) was calculated from the difference in accumulation in control cells (Acontrol) versus lysosome-inactivated cells (Alyso,inact) according to:(2)where Reccontrol and Reclyso,inact are the drug recovery at the end of each experiment, and Nnuc is the average number of counted cell nuclei per image on the simultaneously handled reference plate. The amount in lysosomes was scaled to 200,000 cells per well using fcell no. = 805.14, which relies on the size of the acquired images versus the well size and an average of 1.27 nuclei per hepatocyte (determined separately, data not shown).
Additionally, the percentage of accumulated drug in lysosomes (pacc) of total drug applied in the medium (Atotal) was calculated, as well as the percentage of total cellular content corrected for recovery. The statistical significance of reduced accumulation in lysosome-inactivated cells compared to control cells was tested with a one-sided t test.
Characterization of Rat Hepatocyte Lysosomes.
Freshly isolated rat hepatocytes were plated (36,000 cells/well) in black 96-well Biocoat Collagen I plates (Corning Incorporated) and allowed to adhere for 24 hours at 37°C and 5% CO2. Cells were washed twice with prewarmed PBS. Williams Medium E was added to each well containing 3 µM Lysosensor Yellow/Blue DND-160. After 1 minute of incubation, ratiometric imaging of fluorescence using a Fura-2 filter set on an ImageXpress Micro with 1000 milliseconds of exposure was performed. The calibration was done according to Wang et al. (2017) from pH 4.5 to 6. Images were analyzed in ImageJ (version 1.43u). In short, background was subtracted, regions of interest were identified by finding circular spots with the size of the lysosomes, the fluorescence intensity was measured in both channels, and the ratio of corresponding regions of interests was calculated. The ratio was transformed into a pH value using the calibration curve. The experiment was conducted with hepatocytes from six different rats with a total of >350,000 lysosomes being analyzed.
Prediction of Lysosomal Drug Sequestration.
The intracellular distribution of drugs to lysosomes was calculated based on the Henderson-Hasselbalch equation, experimentally determined hepatocyte-specific parameters, and in silico predicted compound properties. The extent of lysosomal sequestration L was calculated by integrating the partitioning of a drug over the whole range of possible intracellular pH values, see eq. 4 and the simplified schematic in Fig. 3. To weight the contribution of a pH value to the overall accumulation, we adjusted the volume of the lysosomal compartment at a given pH V(pH) based on the experimentally determined pH distribution in hepatocytes, the average lysosomal volume Vlyso, and the total number of lysosomes per cell Nlyso by:(3)where p(⋅) is the probability density function represented by a normal distribution with the mean and variance σ2. A constant pH of 7.2 in the nonacidic compartment pHcyto (Hallifax and Houston, 2012) and a minimum possible pHmin of the lysosome of 4 are assumed. The hepatocyte-specific parameters used in the calculations are listed in Table 2. The Henderson-Hasselbalch–based concentration ratio KL of mono-/dibasic compounds between lysosomes and the surrounding medium is stated in MacIntyre and Cutler (1988a). We incorporate KL into our equation to calculate a lysosomal sequestration extent L for the entire lysosomal system in relation to the cytosolic drug content by:(4)where VB is the volume of the nonacidic compartment in a hepatocyte. The acid dissociation constants of basic drug moieties used to derive the KL value were predicted in silico using ADMET Predictor (Simulations Plus Inc., Lancaster, CA). A ready-to-use Excel file to predict the lysosomal trapping by the aforementioned equations and only requiring the pKa values of compounds can be found in the Supplemental Material.
Results
Indirect Fluorescence-Based Assay for Lysosomotropism.
We chose a fluorescence-based method, which uses the lysosomotropic dye LTR to stain lysosomes, and Hoechst 33342 to stain cell nuclei in order to avoid misinterpretations due to cytotoxicity in plated rat hepatocytes as the cell system. The final LTR concentration of 50 nM was chosen since it gives a strong signal without overloading the cells (pilot experiments; data not shown). The plated rat hepatocytes showed well-defined fluorescent spots, which was abolished by coincubation with 25 mM NH4Cl, indicating a functioning proton gradient in the lysosomes, and thus an intact lysosomal system (Fig. 4). To establish the assay, six well-known reference compounds were chosen, including four lysosomotropic compounds (chloroquine, imipramine, fluoxetine, and propranolol) and two non-lysosomotropic compounds (diclofenac and olaparib). The compounds were selected based on their known lysosomotropic behavior (Lemieux et al., 2004; Nadanaciva et al., 2011; Kazmi et al., 2013) and their physico-chemical properties with basic pKa values spanning from low to high. Additionally, with a log DpH7.5 between 1.6 and 3.2, all reference compounds can be expected to passively cross the lysosomal membrane (Table 3). For developing the evaluation method of the assay, the compounds were measured in a concentration range of 0.5–100 µM. The fluorescence threshold of ≥500 (intensity scale 0–4096) was identified to omit the background and cytosolic fluorescence, allowing the lysosome-specific analysis of LTR, which is different from previously published assays (Kazmi et al., 2013). Due to random attachment of hepatocytes on the plates the number of cells, and therefore the size of the lysosomal system per image, differs. To compensate for this, we calculated the cell number independent parameter plyso/norm as the lysosomal fluorescence (Fig. 5). The concentration dependency of this lysosomal fluorescence was used as the basis for IC50 calculations as a measure of lysosomal trapping. None of the compounds had cytotoxic effects over the concentration range tested. Propranolol, imipramine, chloroquine, and fluoxetine all caused a significant reduction of LTR fluorescence in lysosomes with IC50 values of 15 ± 2.1, 4.8 ± 1.2, 3.9 ± 0.5, and 8.0 ± 2.4 µM, respectively. The negative reference compounds diclofenac and olaparib both did not alter lysosomal fluorescence significantly. The strength of the lysosome-specific fluorescence assessment results in very sensitive (i.e., lower IC50) values and a strong resolution.
Direct Quantification of Lysosomal Drug Content.
Additional investigations were performed with the aim to establish a quantitative link between fluorescence-based IC50 values and the extent of lysosomal drug sequestration. We quantified the intracellular content of selected compounds via LC-MS/MS in control hepatocytes, hepatocytes with inactive lysosomes (+25 µM monensin), and after coincubation with LTR. The results indicate that there is no interference by the presence of 50 nM LTR on the lysosomal drug content as shown for propranolol and imipramine (Fig. 6, A and D). In cells with inactive lysosomes the amount of drug was markedly decreased, showing an effective lysosome inactivation by 25 µM monensin. The reduction is significant (P < 0.05) for imipramine and highly significant (P < 0.01) for propranolol at every concentration tested (Fig. 6, A and D). Assuming complete lysosomal inactivation, the difference between the cellular drug accumulation of control cells and cells with inactive lysosomes represents the extent of sequestration by lysosomes, i.e., the lysosomal drug content. The lysosomal content reaches a plateau at 0.77 and 0.57 nmol per 200,000 cells for propranolol and imipramine, respectively, indicating saturation. Notably, this saturation occurs at concentrations beyond which the cellular drug content continues to rise. While the plateau is similar, the saturation concentrations differ between the compounds with imipramine being more sensitive (Fig. 6, B and E). Furthermore, the distribution to lysosomes can be determined with our assay by expressing the lysosomal drug content relative to the drug in control cells. At 1 µM the endo-/lysosomal system holds about 56% of the cellular propranolol and about 50% of the imipramine, meaning most of the cellular propranolol and imipramine is located in the endo-/lysosomal system. As concentrations increase, lysosomal but not the cellular content becomes saturated causing the relative drug content to decline to 34% and 19% at 100 µM, respectively (Fig. 6, C and F).
Linking the Indirect to the Direct Assay.
As depicted in Fig. 7, A and C, the reduction of lysosomal fluorescence is mirrored by a rising lysosomal drug content for imipramine and propranolol, respectively. While the lysosomal content rises to a maximum, the lysosomal fluorescence approaches a minimum fluorescence level. When the lysosomal drug content is expressed as the proportion of applied drug it completely overlays with the lysosomal fluorescence reduction (Fig. 7, B and D). Consequently, the IC50 values of both assays are similar, with 4.8 ± 1.2 µM versus 4.9 ± 0.9 µM for imipramine and 15 ± 2.1 µM versus 9.5 ± 0.9 µM for propranolol. In addition, the very similar Hill slopes indicate a direct relation between lysosomal drug content and displaced LTR fluorescence over the whole concentration range. Therefore, an estimation of lysosomal drug accumulation and its concentration dependence can be derived from the fluorescence data.
Lysosome Characterization in Cultured Rat Hepatocytes.
The endo-/lysosomal volume and the pH value are the most important cellular parameters for lysosomal sequestration. Therefore, we determined both parameters experimentally for the hepatocytes used in our assays. The size of hepatocytes was determined with bright-field microscopy (465 analyzed), and for the endo-/lysosomal volume we carried out LTR staining followed by counting the fluorescence spots (5143 counted) and determining their area (3200 analyzed) to calculate a spherical volume (data not shown). The proportion of the endo-/lysosomal system in plated rat hepatocytes is 1.49% ± 0.08% of the cellular volume. The lysosomal pH was determined with the pH-sensitive dye Lysosensor Yellow/Blue DND-160, which changes its fluorescence spectrum depending on the surrounding pH. A calibration curve was constructed from pH 4.5 to 6.0 and extrapolated for pH values outside of these margins. Hepatocytes from six individual rats were subject to investigation with >350,000 endo-/lysosomes analyzed in total for sufficient statistical power. The endo-/lysosomal pH values found cover a wide range from pH 4 up to 7.2 with different frequencies of occurrence that follow a Gaussian distribution with a mean at pH 5.53 (Fig. 8). See the Supplemental Material for a detailed distribution.
In Silico Prediction of Lysosomal Sequestration.
In addition to the experimental data, the lysosomal accumulation was calculated according to eq. 4 for propranolol and imipramine. Both compounds are strong bases, with a pKa value of 9.4 for propranolol and two basic pKa values of 9 and 2.2 for imipramine. However, the second basic pKa value of imipramine does not have a significant influence on the pH-driven accumulation, thus its behavior is effectively monobasic. The calculated amounts in the lysosome (L) for propranolol and imipramine are 1.15 and 1.14 times the amount in cytosol, respectively, i.e., the endo-/lysosomal system holds about 53% of the intracellular drug content (see the Supplemental Material). This prediction matches perfectly with the direct quantification assay at low concentrations where the lysosomes are not saturated (see the horizontal line in Fig. 6, C and F).
Discussion
Understanding intracellular distribution and intracellular concentrations of drugs is a topic of growing interest, as highlighted in the International Transporter Consortium white paper of Guo et al. (2018). The implications of intracellular pharmacokinetics in safety and efficacy make early information crucial for the selection of new drug candidates. The subcellular distribution to lysosomes plays a major role for basic lipophilic compounds. While there are some assay formats to assess lysosomotropism, none completely fulfills the requirements needed in drug discovery. For instance, the assay by Kazmi et al. (2013) requires compound concentrations up to 500 µM, which most often cannot be achieved in the lead optimization phase. Nadanaciva et al. (2011) determined lysosomal trapping in a cell line that seems overly sensitive to cytotoxicity, thereby biasing the results. The fluorescence assay presented herein overcomes these shortcomings and can also be quantitatively linked to the lysosomal drug content. As the cell system, we used freshly isolated rat hepatocytes, which are rich in lysosomes and are routinely available in drug metabolism and pharmacokinetic departments (Reichel and Lienau, 2016). The assay principle, however, is also applicable to other palatable cells.
Indirect Fluorescence-Based Assay for Lysosomotropism.
In analogy with previously published assays (Lemieux et al., 2004; Nadanaciva et al., 2011; Kazmi et al., 2013), we used LTR as a fluorescence marker, however, modifying the analysis by specifically quantifying the fluorescence associated with lysosomes. This is the major difference from Kazmi et al. (2013), who measured the fluorescence irrespective of its cellular location. Using a set of reference compounds we have shown that this modification reduces the IC50 values, and therefore the concentration range to be tested (e.g., for chloroquine from 220 to 4 µM, for propranolol from 80 to 15 µM, and for imipramine from 260 to 5 µM) compared with Kazmi et al. (2013). The more sensitive method now allows one to also measure low soluble discovery compounds. When measured on an automatic microscope as used by Nadanaciva et al. (2011) this method is amenable for higher throughput.
To avoid misinterpretations, we simultaneously monitored cell viability in the assay. None of the compounds showed cytotoxic effects at any concentration tested, attesting to the suitability of rat hepatocytes.
Direct Quantification of Lysosomal Drug Content.
To obtain better understanding of the meaning and implications of the IC50 values from fluorescence assay, we directly determined the drug content in the lysosomes. To quantify the lysosomal drug content, we directly measured the drugs via LC-MS/MS in hepatocytes with intact and inactive lysosomes, which allowed quantifying both the lysosomal and the total drug content in the cells. To abolish lysosomal trapping, we used the ionophore monensin, which collapses the pH (proton) gradient between the lysosome and cytosol. Therefore, it specifically disables the pH-driven part of drug accumulation, leaving other distribution processes such as lipid and protein binding intact.
At low compound concentrations lysosomes contain the highest fraction of total intracellular drug with about 50% and 56% for imipramine and propranolol, respectively. This is comparable to the data obtained by Kazmi et al. (2013) despite using different cells and ammonium chloride to inhibit lysosomotropism, whereas the extent of lysosomal trapping of both compounds is very similar; their IC50 values differ by a factor of 3 (as previously shown). This suggests they carry information that is more related to the concentration dependence of the lysosomal trapping rather than the extent.
Linking the Indirect to the Direct Assay.
With the possibility of drug-drug interactions involving lysosomes (Daniel and Wójcikowski, 1999a,b; Daniel, 2003), we first ensured that LTR at the concentration used is not affecting the sequestration of test drugs in the fluorescence assay. The results shown in Fig. 6, A and D, confirm that the presence of LTR does not affect lysosomal accumulation of test drugs, a prerequisite to a quantitative link between the assays. Figure 7, A and C, clearly suggests a relationship between the two assays, i.e., the declining lysosomal fluorescence with increasing drug concentrations is directly mirrored by the rising lysosomal drug content with both readouts approaching a plateau at high concentrations. Expressing the lysosomal drug content in relation to the amount of drug applied in the assay, a direct relationship with the fluorescence assay is obtained (Fig. 7, B and D). Remarkably, there is a complete overlay of the data points from the two assays for both drugs, including Hill slopes and IC50 values. This implies a quantitative displacement of LTR by both drugs, and therefore the reduction of lysosomal fluorescence can be used to describe their concentration-dependent lysosomal accumulation. This thus allows one to reduce the number of measurements over the whole concentration range in the direct assay down to just the highest compound concentration as the anchor point. This simplification is possible only through the lysosome-specific evaluation of the fluorescence in the indirect assay, such that both assays can now operate in the same concentration range.
In Silico Prediction of the Extent of Lysosomal Trapping.
For an accurate in silico prediction of lysosomal trapping, comprehensive information on the acidity and the fractional volume of the endo-/lysosomal system is needed due to tremendous variations between cell types ranging from 0.23% up to 7.8% of the cellular volume (de Duve et al., 1974; Blouin et al., 1977; MacIntyre and Cutler, 1988a; Ufuk et al., 2015) with pH values reported in the range of 4.65–5.18 (Regec et al., 1989; Tietz et al., 1990; Kharbanda et al., 1997). We have, for the first time, determined a full profile of the pH values and corresponding fractional volumes of the endo-/lysosomal system in rat hepatocytes (Fig. 8). Simply using a mean lysosomal pH would underpredict the extent of lysosomal sequestration because the lysosome/cytosol concentration ratio (KL) rises exponentially with decreasing lysosomal pH (Fig. 8). However, due to insufficient knowledge on the pH distribution of lysosomes this oversimplification was frequently practice until now. The improvement in the predictivity by using the full pH profile is shown through the following examples. For two monobasic drugs with pKa values of 7 and 9 the predicted extents are 31% (instead of 23%) and 53% (instead of 43%), respectively, calculated by eq. 4 (see the Supplemental Material) versus the mean lysosomal pH (Table 2). The superiority of the prediction is confirmed by the experimental data on imipramine and propranolol (both pKa ≈ 9), resulting in values of 50% and 56%, respectively.
However, it needs to be kept in mind that these predictions are valid only for low concentrations where lysosomal trapping is not saturated because this may change the lysosomal pH profile or volume. The concentrations at which saturation will occur have to be determined experimentally, e.g., via the IC50 values of the indirect assay. The in silico predictions for a given drug are only valid up to concentrations of its lysosomal IC50. If information on the extent of trapping beyond these concentrations is needed this can be obtained by the direct assay (Fig. 6, C and F).
Application in Drug Discovery.
Lysosomal drug accumulation, although likely to occur for most basic lipophilic drugs, is not regarded as a property requiring routine screening in drug discovery. It is rather a subject of special investigation that is indicated by certain trigger points, for example, 1) to elucidate unusual compound accumulation in (lysosomal-rich) tissues observed in preclinical animal species requiring a more mechanistic understanding of the mechanism and/or handle for optimization; 2) to direct the distribution of compounds whose on-targets or off-targets are within lysosomes in or out of the organelle; 3) to specifically modify the lysosomotropism of compounds in differentiation from a front-runner project or competitor drugs; 4) to reduce the propensity of compounds to induce lysosomal sequestration-related phospholipidosis; or 5) to examine whether a saturable uptake of a basic compound in hepatocytes is due to saturation of an uptake transporter or just the result of lysosomal trapping that may also get saturated, to name but a few. While some of these questions require mechanistic studies to identify whether lysosomal trapping is involved, other questions may need compound screening, rank-ordering, and optimization at a higher throughput. The set of assays proposed herein is principally able to meet these requirements, allowing a tailored application in the project work.
In conclusion, we have established an indirect fluorescence and a direct quantification assay set that can be quantitatively linked to assess the susceptibility, concentration dependency, and maximal extent of lysosomal trapping in a drug discovery context suitable for either screening or more mechanistic examinations. In addition, for the first time we have determined the full endo-/lysosomal pH profile of cultured rat hepatocytes, now allowing a more accurate in silico prediction of the extent of lysosomal trapping of basic lipophilic compounds.
Acknowledgments
We thank Stephan Menz for assistance with mathematical questions and Daniel Trost and Martin Klein for support in hepatocyte isolation and LC-MS/MS quantification.
Authorship Contributions
Participated in research design: Schmitt, Lienau, Reichel, Fricker.
Conducted experiments: Schmitt.
Performed data analysis: Schmitt, Lienau.
Wrote or contributed to the writing of the manuscript: Schmitt, Lienau, Reichel, Fricker.
Footnotes
- Received September 17, 2018.
- Accepted October 22, 2018.
This work was financed by Bayer AG, Germany.
↵This article has supplemental material available at dmd.aspetjournals.org.
Abbreviations
- DAPI
- -4′,6-diamidino-2-phenylindole
- LC-MS/MS
- liquid chromatography–tandem mass spectrometry
- LTR
- LysoTracker Red DND-99
- PBS
- phosphate-buffered saline
- Copyright © 2018 by The American Society for Pharmacology and Experimental Therapeutics