Molecular Design of a Pathogen Activated, Self-Assembling Mechanopharmaceutical Device

Abstract

Weakly basic small molecule drugs like clofazimine can be used as building blocks for endowing cells with unnatural structural and functional elements. Here, we describe how clofazimine represents a first-in-class mechanopharmaceutical device, serving to construct inert, inactive and stimulus responsive drug depots within the endophagolysosomal compartment of cells of living organisms. Upon oral administration, clofazimine molecules self-assemble into stable, membrane-bound, crystal-like drug inclusions (CLDI) that accumulate within macrophages to form a “smart” biocompatible, pathogen activatable mechanopharmaceutical device. Upon perturbation of the mechanism maintaining pH and ion homeostasis of these CLDIs, the inert encapsulated drug precipitates are destabilized, releasing bioactive drug molecules into the cell and its surrounding. The resulting increase in clofazimine solubility activates this broad-spectrum antimicrobial, antiparasitic, antiviral or cytotoxic agent within the infected macrophage. We present a general, molecular design strategy for using clofazimine and other small molecule building blocks for the cytoplasmic construction of mechanopharmaceutical devices, aimed at rapid deployment during infectious disease outbreaks, for the purpose of pandemic prevention.

1. Introduction

To facilitate cytoplasmic construction, many small molecule chemical agents are known to preferentially bioaccumulate within target organelles or specific cells, such as macrophages [1, 2]. If concentrated inside organelles at high concentrations, these poorly soluble organic molecules will exceed their solubility in the local microenvironments, forming insoluble precipitates [3-6]. As this soluble-to-insoluble phase transition occurs, precipitated molecules may adopt a disordered liquid or aggregated state, or they can self-assemble into ordered supramolecular structures such as liquid crystals or solid-state, three-dimensional crystals [7].

Taking advantage of this natural targeting phenomenon, organic small molecules could be designed and screened for their ability to create self-assembling, mechanopharmaceutical devices possessing unique physical, chemical, and biological features [8-12]. Due to the distinct properties of ordered molecular structures, self-assembled mechanopharmaceutical devices can be readily retrieved, isolated, or monitored in living organisms [13]. The cytoplasmic construction of mechanopharmaceutical devices has already been demonstrated as a strategy to endow cells of living organisms with functional supramolecular structures possessing unique fluorescence and photoacoustic signals, as well as a means to massively load cells for drug targeting and delivery purposes, and for measuring the phenotypic properties of the targeted cells [14].

However, perhaps the most clinically relevant application of a self-assembling mechanopharmaceutical device is the possibility of creating a self-assembling, pathogen activatable drug depot. Following oral drug administration, the absorbed drug molecules circulate in the blood and partition within macrophages, where they self-organize into supramolecular structures that function as biocompatible, intracellular reservoirs of precipitated drug [15]. Following drug loading during a prolonged oral dosing period, sustained pharmacodynamic activity can be maintained through drug elution from the self-assembled drug reservoir, which is reflected in an increasing systemic half-life of the drug when administered in long-term treatment regimens [16].

2. Clofazimine as a Self-Assembling Biomolecule

2.1. Solubility and Lysosomal Targeting

One of the most studied pharmaceuticals to undergo a soluble-to-insoluble phase transition inside the organism following oral administration in therapeutically relevant conditions is the FDA approved drug clofazimine (CFZ). CFZ has cured over 14 million leprosy patients and is currently recommended by the WHO as part of the standard treatment of leprosy and MDR-TB [17].

Due to limited aqueous solubility (<0.01mg/L) [17] and high lipophilicity (logP=7.66), CFZ accumulates in fatty tissue during a short-term treatment regimen [18]. Concomitantly, the ionizable amine group (apparent pKa = 6.08) also leads to drug accumulation in acidic organelles such as the lysosome (pH 4.5-5.0). In both human and animal models, it has been shown that CFZ forms insoluble Crystal-Like Drug Inclusions (CLDIs) through ion-trapping and supramolecular complexation of the protonated monocationic form of the weakly basic drug, within the macrophage lysosome [19]. As a result of this soluble-to-insoluble phase transition, CFZ exhibits an increasing volume of distribution following a prolonged oral dosing period, which leads to the accumulation of drug in organs of the reticuloendothelial system to much greater extent than would be predicted by a simple partitioning into body fat. Thus, the continuously increasing volume of distribution, as a function of intracellular drug crystal deposition, complicates traditional compartmental analysis of CFZ that stipulate a constant volume of distribution for pharmacokinetic parameter estimates [20, 21].

Due to the pKa of the amine functional group of CFZ, acting together with the acidic, lumenal pH of the lysosomes, CFZ becomes trapped inside lysosomes as the conjugate acid CFZH+. The concentration of chloride ions within the lysosomal compartment is higher than that of other anions, such that CFZ is primarily present as a hydrochloride salt (CFZH⁺Cl⁻). Being the most concentrated anion in the body, chloride is typically present at 10-to-100-fold greater concentration than any other anion (Figure 1A) [22, 23]. Due to the common ion effect arising from the high chloride concentration in lysosomes, the hydrochloride salt is markedly less soluble than the free base within the lysosome (CFZH⁺Cl⁻ lysosomal solubility = 3 nM vs CFZ solubility = 480 nM), lending itself to rapid precipitation with continued drug loading. Combined with significant protein binding, the vastly lower solubility of the lysosomal salt form of the drug becomes the primary determinant of the freely soluble CFZ concentrations that circulate in the blood and partition in the cells throughout the organism, allowing the drug to remain well below its active concentration. Thus, CFZ is well-tolerated despite its promiscuity.

Figure 1.

Lysosomal acidification machinery determines CFZ solubility. A) Physiologically relevant anion concentrations in the interstitial fluid, cytosol, and lysosome. B) Cell model with relevant compartmental pH and molecular structures. C) pH-Dependent solubility curve, i) common-ion effect. ii) pH<pHmax, iii) pH>pHmax. Solubility of CFZH+Cl− salt form and CFZ free base with respect to pH change generated using literature reported solubility parameters (S₀=480 nM, pHmax=4.5, and K_sp=332.3 μM²). The dashed line, solid blue line, and solid black line segments represent the solubility of the salt form with or without the common ion effects, and the solubility of the freebase respectively. Using established equations, total drug solubility with respect to pH change was created in RStudio Version 1.2.1335. The data points were calculated based on equations listed above for corresponding pH ranges. The embedded packages dplyr, coin, and ggplot2 were used to clean the calculated data and make the plot.

* Thyroid tissue only

The naturally acidic lysosome of macrophages contrasts with the slightly basic internal environment of the cytosol (pH 7.0-7.4) due to a high density of vacuolar-type H+-ATPase (V-ATPase) pumps on its membranes, which use the free energy of ATP hydrolysis to pump protons into the lumen of the lysosome [24-26]. The acidic environment of the lysosome plays an important role in the decomposition of nutrients and pathogenic organisms following phagocytosis. Although lysosomes exhibit specialized secretory functions in some cell types [27], the expression level of V-ATPase is relatively universal. Because the V-ATPase activity separates electric charge and generates a transmembrane voltage, another ion must move to dissipate this voltage for net pumping to occur. This counterion may be either a cation (moving out of the lysosome) or an anion (moving into the lysosome) [26]. Evidence points to chloride as this counterion using a lysosomal chloride channel (ClC-3). As such, the inhibition of V-ATPase would lead to significant change in pH and chloride ion concentration [28]. Thus, as the main counterion in the lysosome, chloride ion concentrations are relatively high, around 110mM [22], which also serves to drive the precipitation of CFZH⁺Cl⁻.

The master regulator of lysosomal biogenesis, Transcription Factor EB (TFEB), is actively expressed in macrophages and its overexpression is an important determinant of macrophage’s phagolysosomal function [29-31]. The expanded phagolysosomal compartment of macrophages is ideally suited to maintain low CFZ solubility, through the overexpression of V-ATPase and ClC-3, maintaining high proton and chloride concentrations upon CFZH⁺Cl⁻ precipitation (Figure 1B). Macrophages are thus equipped with the ability to maintain the acidic environment in the lysosomes concomitantly with the degradation of large amounts of proteins as may be present in phagocytosed bacteria [32]. Lysosomes are known to contain more than 60 different hydrolytic enzymes which require a low pH to conduct optimal catalytic activity. The macrophages therefore are especially equipped to maintain a pH gradient to compartmentalize these degradative enzymes within the phagolysosomal compartment while protecting the cytosol and the rest of the cell from their catalytic activity.

Lysosomal biogenesis, driven by TFEB, is upregulated upon CFZ administration, increasing the expression of resident lysosomal proteins in the cell [33, 34]. Upon phagocytosis of CLDIs, Akt phosphorylation is induced, likely contributing to TFEB activation and resistance to apoptosis [15]. It is possible that cytoplasmic construction in macrophage lysosomes benefits from specific transcriptional responses to drug accumulation through the activation of TFEB. The respective increase in lysosomal biogenesis can control the expansion of the lysosomal compartment and allows for accommodation of incremental increase in drug loads. TFEB activation expectedly leads to increased expression of the V-ATPase, which reduces lysosomal stress, maintaining the low pH, and actively allowing for increased intracellular drug sequestration.

Based on the Law of Mass Action, the high lysosomal chloride concentration determines the low solubility of the protonated form of CFZ according to the salt solubility product (K_sp) of CFZH⁺Cl⁻ (Eq. 1) [14, 35].

$${\mathrm{K}}_{\text{sp}}=[{\text{CFZH}}^{+}{]}_{\mathrm{S}}\times [{\text{Cl}}^{-}]$$ Eq. (1)

As such, although the solubility of the protonated CFZ in pure water is much greater than the solubility of the unprotonated, conjugate base (Figure 1C), the high proton and chloride ion concentrations within the lysosome ensure that the solubility of the insoluble salt complex is lower than the intrinsic free base solubility. In the lysosomes, the lumenal microenvironment contains chloride ions at a high concentration, well above the 18 μM concentration of chloride ions that is present when the crystalline, hydrochloride salt form of the drug is dissolved in water at or below the pHmax, given the 1:1 stoichiometric relationship of CFZH⁺ and Cl⁻ in CFZH⁺Cl⁻. In the lysosomes, the solubility of CFZH⁺ at or below the pHmax is expectedly governed by the chloride concentration, decreasing the total solubility according to the common ion effect (Eq. 2; Figure 1C; region i).

$${\mathrm{S}}_{\mathrm{T}}=\frac{Ksp}{C{l}^{{}^{-}}}\times (1+{10}^{\text{pH}-\text{pKa}})$$ Eq. (2)

In pure water, when the pH is at the pH_max, the soluble CFZH⁺ exists in equilibrium with the thermodynamically more stable solid form of the hydrochloride salt (Figure 1C; region ii), its concentration is determined according to the following equation:

$${\mathrm{S}}_{\mathrm{T}}=\sqrt{\text{Ksp}}\times (1+{10}^{\text{pH}-\text{pKa}})$$ Eq. (3)

Lastly, when pH> pH_max, the total drug solubility in water is driven by the intrinsic solubility of the free base (Eq. 4).

$${\mathrm{S}}_{\mathrm{T}}={\mathrm{S}}_0\times (1+{10}^{\text{pKa}-\text{pH}})$$ Eq. (4)

As the pH is increased above the pHmax by adding base (e.g., NaOH) to a pure aqueous solution of CFZH⁺Cl⁻, the ratio of unionized to ionized species in solution will increase due to Henderson-Hasselbalch equilibrium. As the pH becomes increasingly alkaline, the total solubility of CFZ molecules will gradually decrease, approaching the intrinsic aqueous solubility of the free base, 480 nM (Figure 1C; region iii). Nevertheless, in a lysosome containing CLDIs, a rise in pH above the pHmax decreases the ratio of protonated to unprotonated CFZ, increasing the total solubility of the drug accompanying the release of precipitated drug sequestered as insoluble CFZH⁺Cl⁻ in the CLDI in a manner that is also dependent on the membrane permeability properties of the solubilized drug molecules.

2.2. Supramolecular Assembly of a Mechanopharmaceutical Device

With sustained extracellular concentrations of CFZ, ion-trapping and thermodynamic stability of the hydrochloride salt leads to supersaturation and precipitation within the lysosome. The supersaturation of CFZH⁺Cl⁻ initiates the formation of disordered supramolecular drug complexes that eventually become organized as a three dimensional crystalline solid. CFZ is highly lipid soluble, allowing it to partition within and across the cellular membranes, achieving its pH-dependent equilibrium between the unprotonated and protonated base in the aqueous lumen of the various subcellular compartments. Nevertheless, the primary determinant of intracellular drug mass disposition is the low hydrochloride salt solubility of 3 nM in the lysosomes, which drives drug precipitation within the lysosome (Eq. 1). Data from transmission electron microscopy, deep etch freeze fracture electron microscopy, powder X-ray diffraction and Raman spectroscopy all point to the drug molecules present in CLDIs as a thermodynamically stable, solid 3D structure (Figure 2A-C). Based on in vitro experiments, the crystalline stability is due to the precise structure of the crystalline drug salt, and not simply due to the high logP [14, 35]. As the drug precipitates, grows and becomes organized within the macrophage lysosome, membranous lipid molecules associated with the surface of crystalline drug complexes to form the supramolecular, membrane bound CLDI complex (Figure 2D, E). Multiple intracellular deposits of CLDI structures can develop, leading to a prominent repository of self-assembled CLDIs within the cytoplasm (Figure 2F).

Figure 2.

Image sequence for supramolecular CLDI organization. A) CFZH⁺Cl⁻ structure in ChemDraw B) CFZH⁺Cl⁻, 3D-molecular orientation in Mercury CCDC with colors corresponding to the colored functional groups in part A. C) Supramolecular crystalline organization in Mercury CCDC. D) Freeze fracture image of a macrophage sequestered intracellular drug inclusion [19]. E) Freeze fracture image of a membrane bound crystal formation [19]. F) Multiple CLDIs forming inside a macrophage [19].

Despite the proportionally large volume taken up by intracellular CLDIs, macrophages loaded with CFZ remain alive and functional. While CFZ is expectedly bioactive at the concentration of the intrinsic free base solubility limit, the crystallized form of CFZH⁺Cl⁻ within CLDIs is biologically inactive against most of its known molecular targets, as its solubility is (at least) two order of magnitudes lower than the concentrations required for activity [15]. The difference in reactivity between the CLDI form of the drug and the freely soluble neutral form is an essential quality of this mechanopharmaceutical device granting CLDIs the ability to lie dormant within the cell until activated by an external stimulus.

2.3. Macrophages as Endogenous Tools of Self-Assembly

In addition to the essential role played by lysosomal proton pumps and chloride transporters present on the surface of the bounding membrane of macrophage lysosomes, macrophages can alter their internal membrane architecture to maximize the volume of substances sequestered within the lysosome, referred to as a “phagolysosomal cargo”. Continued accumulation of cargo has been associated with increased formation of larger volume vesicles and to the disappearance of smaller vesicles. It can be inferred that by expanding the volume of intracellular vesicles, macrophages maximize the available cargo capacity while minimizing the overall membrane surface area delimiting the cargo compartments [36, 37]. This allows for increased cargo accumulation within the cell, and consequently, an increase in volume of distribution of a bioaccumulating, precipitating drug.

In terms of the pharmacokinetics properties that allows for macrophage dependent self-assembly, CFZ is the only known drug that has been extensively studied in this regard. Measurements of the entire cell population of liver, lung, and spleen indicate an active macrophage response to the increasing drug cargo load. This biological response includes an increase in the number of macrophages occurring in parallel to changes in intracellular membrane organization that accumulates the self-assembled mechanopharmaceutical CLDI devices. In pathological states an expansion in the macrophage population is typically associated with a pro-inflammatory response [38, 39]. However, the accumulation of CFZ and its self-assembly in macrophages was accompanied by the activation of anti-inflammatory signaling pathways and was not associated with obvious toxicological manifestations [40, 41]. In addition, cargo loading of macrophages induced increases in the number of cells and accompanying increases in organ mass, resulting in splenomegaly following extensive drug loading [34]. Additionally, changes in histological organization have been observed, such as large clusters of macrophages called granulomas [34, 42, 43]. These changes enable a spatially localized, collective intracellular cargo load associated with the self-assembly of this functional, mechanopharmaceutical device.

2.4. Pharmacokinetics of Clofazimine

The macrophage targeted accumulation of CLDIs expectedly alters the whole-body pharmacokinetics in a context-dependent manner. As drug sequesters within lysosomal compartments of macrophage cells, the volume of distribution continues to expand nonlinearly. Due to the dissemination of macrophages throughout the tissues of the body, widespread accumulation of CFZ in both free base and CLDI form have been observed in murine and human tissues (Table 1).

Table 1.

CFZ serum and tissue concentration data under variable infections and dosing regimens. [16, 34, 44-48]

CFZ Dose	Dosing Duration	Infection Status	Number of Subjects	Peak Serum Concentration (μM)	Spleen Concentration (μM)	Lung Concentration (μM)	Liver Concentration (μM)	Ref
Human Serum and Tissue Concentrations
200 mg SD	Single Dose	Healthy	8	0.861 +/− 0.289	N/A	N/A	N/A	34
300 mg daily for 1-2 weeks, 200 mg for 2-8 weeks	2 Months	Drug Resistant TB (PROBeX)	79	0.766 (0.340-1.452)	N/A	N/A	N/A	44
200 mg	14 Days	Drug Sensitive Tuberculosis	60	0.460 (0.145-0.871)	N/A	N/A	N/A	44
100 mg Three times per week	N/A	Leprosy	N/A	1.06	N/A	N/A	N/A	45
100 mg daily	N/A	Leprosy	N/A	1.48	N/A	N/A	N/A	45
300 mg daily	N/A	Leprosy	N/A	2.11	N/A	N/A	N/A	45
400 mg daily	N/A	Leprosy	N/A	2.95	N/A	N/A	N/A	45
High-Dose daily	Long-term	Leprosy	1	N/A	2532	N/A	443	46
100 - 300 mg daily	35 days to 243 days	Leprosy	3	N/A	2638 (1266 - 4009)	2251 (1266 - 2954)	3517 (1899 - 6752)	47
Mice Serum and Tissue Concentrations
25 mg/kg	Single Dose	Healthy	5	0.91	3.65	1.6	5.87	48
25 mg/kg	1 week	Healthy	3	1.94 +/− 0.63	17.11 +/− 6.58	11.04 +/− 5.42	14.88 +/− 3.90	48
25 mg/kg	2 weeks	Healthy	3	1.79 +/− 0.063	23.59 +/− 1.10	23.78 +/− 5.84	23.74 +/− 0.61	48
25 mg/kg	4 weeks	Healthy	3	2.22 +/− 0.084	106.41 +/− 4.05	63.26 +/− 10.55	70.16 +/− 13.06	48
25 mg/kg	8 weeks	Healthy	3	4.75 +/− 0.063	402.46 +/− 203.40	110.96 +/− 14.22	126.54 +/− 56.02	48
25 mg/kg	12 weeks	Healthy	3	2.74 +/− 0.61	9884.67 +/− 864.49	86.15 +/− 17.22	173.80 +/− 17.75	48
25 mg/kg	16 weeks	Healthy	3	2.55 +/− 0.44	7945.84 +/− 1502.87	62.79 +/− 27.58	234.82 +/− 163.10	48
25 mg/kg	20 weeks	Healthy	3	2.11 +/− 0.25	14292.4 +/− 1483.6	78.89 +/− 11.50	226.21 +/− 26.69	48
36 mg/kg	3 weeks	Healthy	6	16.96 +/− 0.422*	2012.94 +/− 1446.43	N/A	675.2 +/− 222.31	16
36 mg/kg	8 weeks	Healthy	6	4.54 +/− 1.88*	27900 +/− 3129.64	N/A	10128 +/− 1126.42	16

^*Plasma concentrations

As can be noted in both species, the concentration of CFZ in tissue is greater than the concentration of CFZ in serum. In fact, while the tissue concentration increases with continual drug load, the serum concentration remains static in the low micromolar range. This discrepancy between drug in soluble aqueous environment and solid tissue is explained by the presence of insoluble drug precipitates within tissue-resident macrophages. This is further supported by microscopic evaluation of tissue samples in both humans and mice. Autopsy and sputum data has revealed CLDIs dispersed throughout multiple tissues in humans after prolonged treatment at a therapeutic dose, indicating that the observed phenomenon is relevant to the human condition, in an FDA-approved, clinical dosing regimen [49-52].

The deviation of the measured CFZ drug concentrations from the expected concentrations predicted by mathematical models used in standard pharmacokinetic analysis has made it difficult to determine standardized therapeutic regimens of CFZ for established indications. CFZ’s pharmacokinetics are further complicated by the difficulty conducting therapeutic drug monitoring due to the discrepancy between blood concentrations and the respective tissue concentrations of interest, such as attempting to monitor lung concentrations in tuberculosis using serum concentrations as a surrogate marker. Nevertheless, the drug is extensively used throughout the world as a treatment for leprosy, multidrug resistant tuberculosis, and nontuberculous mycobacterial infections.

3. Chemical Kinetics of an Intracellular Pathogen Activatable Drug Depot

Upon prolonged oral administration, intracellular CLDIs self-organize into inert drug depots disseminated throughout the macrophage population of the host organism. After macrophages have been systemically loaded, crystallized CFZ remains compartmentalized within intracellular CLDIs, awaiting activation and solubilization. Interference with lysosomal acidification mechanisms, macrophage death, or lysosomal membrane permeabilization are all potential mechanisms by which the thermodynamically stable CLDIs can destabilize and begin to elute soluble drug into the cytoplasm and the surrounding extracellular space. This destabilization of the CLDI initiates a conversion between inactive CFZH⁺Cl⁻ molecules present in CLDIs to bioactive CFZ free base. CLDIs thus represent an unnatural, stimulus responsive drug depot system that is integrated into our body’s natural defense system to combat an emergent infection.

3.1. Mechanisms of CLDI Destabilization

Established studies have shown evidence of lysosomal pH destabilization induced by a variety of infectious pathogenic organisms (Table 2). The same lysosomal destabilization mechanism that leads to the intracellular survival of pathogens can serve as the mechanism that triggers CLDI activation. It has been argued that the effect of a pathogenic organism on the ability of lysosomes to maintain acidic pH plays a role in the stark difference in the ability of mice to form CLDIs depending on their M. tuberculosis infection state [21]. As M. tuberculosis has been shown to arrest the acidification mechanism of lysosomes, the presence of infection may also prevent CLDI formation. When given to healthy patients, administered CFZ would be most prone to accumulate in CLDI form, warranting its consideration as a prophylactic treatment [53]. Because many other pathogens have been shown to destabilize the lysosome (Table 2), this phenomenon is more broadly applicable than to mycobacteria alone. For example, Listeria monocytogenes has been shown to prevent fusion with the lysosome through perforation of the phagosomal membrane using Listeria Lysin O. These perforations allow both pathogen escape and prevent the acidification of the compartment, which is necessary for lysosomal membrane fusion [54]. Similarly, Toxoplasma gondii block phagosome acidification, also preventing membrane fusion. Drug molecules that accumulate in endocytosed vesicles could also provide a parallel lysosomal accumulation pathway [55].

Table 2.

Pathogen Activated Lysosomal Acidification. [53, 56-58, 59-63]

Pathogen	Mean Initial pH	Mean Final pH	ΔpH	Hypothesized Mode of Deacidification	Reference
Interference with Lysosomal Acidification Machinery
Mycobacterium Tuberculosis	4.8	>5.8	>1	Interference with V-ATPase	[53]
β-Coronavirus	4.7	5.7	1	Indirect or direct perturbations in proton pump	[59]
Vibrio parahaemolyticus	*Unquantified Increase			VopQ interference with acidification	[60]
Streptococcus Oralis	*Unquantified Increase			H2O2 production	[61]
Legionella pneumophilia	^Unquantified Increase			V-ATPase inhibition with SidK	[62]
Candida albicans	4.5	7.3	2.8	Phagolysosome alkalinization through NH3 production	[63]
Pathogen Induced Lysosomal Permeabilization
Listeria Myocytogenes	Unquantified Increase			Host cell lysosome permeabilization with Lysteria Lysin O	[56]
Escherichia Coli	Unquantified Increase			Lysosomal permeabilization in type 1-fimbriated E. coli	[57]
HIV	Unquantified Increase			lysosomal permeabilization and cathepsin release	[58]

^*LysoTracker, acidotropic fluorescent probe showed increase in pH

^{^}SidK/V-ATPase biochemical study

In a broader sense, the lysosome as a target for drug accumulation and release could be expanded not only to pathogens which directly deacidify the compartment, but also to those that disrupt the ability of macrophages to maintain ion homeostasis of cytoplasmic organelles. Several pathogens adopt lysosomal membrane permeabilization as a strategy for escape, such as Listeria monocytogenes, Escherichia coli, and human immunodeficiency virus (HIV) [56-58]. Although the universe of pathogens that interfere with lysosomal acidification, and could thus activate CLDIs, is diverse, large, and not fully known, we can illustrate the potential breadth of this approach with a handful of pathogens and mechanisms implicated in lysosomal deacidification (Table 2). The increase in lysosomal pH or otherwise inhibiting lysosomal function is a general, pathogen survival strategy. As macrophage lysosomes are key for our innate immune system to break down invading bacteria and viruses, it comes as no surprise that these invaders have evolved methods of lysosomal deacidification as a common pathway to weaken our defenses.

While it has been shown that prolonged dosing of CFZ results in accumulation and nonlinear pharmacokinetics [20, 21], the same mechanism could be exploited as an avenue for developing prophylaxis against many kinds of pathogens, especially when used in combination with other antimicrobial or antiviral drugs. The pathogen-activated mechanism for drug release from the lysosome is a promising lead into the development and use of other small molecule weakly basic drugs which accumulate in the lysosome as additional components of a pathogen-activatable mechanopharmaceutical device. Upon infection by M. Tuberculosis, which has already shown susceptibility to CFZ, the lysosomal microenvironment becomes perturbed, reducing the quantity of CFZ sequestered into the lysosomal compartment. By the same mechanism, pre-loading cells with CLDIs would also be protective against infection, and other drugs could be administered if any symptoms develop.

3.2. Computational Model of Mechanopharmaceutical Device Activation

To illustrate the kinetics of CFZ solubilization after drug sequestration, a single cell chemical kinetic model was developed in VCell (awillmer: Drug Eluting Device). Utilizing the measured solubility of CFZ free base (480 nM) and the calculated Ksp of the hydrochloride salt (3 nM), time dependent flux of drug leaving the lysosome was predicted under various fixed extracellular and lysosomal concentrations. Three unique lysosomal pHs were chosen to emulate destabilization conditions experienced by exogenous pathogens. At lysosomal pH of 4.5, normal physiological conditions are experienced, leading to CFZH⁺Cl⁻ as the predominate form. As a result, solubility of CFZ in the lysosome will be governed by the Ksp of CFZH⁺Cl⁻ (3nM) in the presence of high lysosomal chloride concentration. Following lysosomal destabilization through a transmembrane proton leak or V-ATPase inhibition, lysosomal pH was increased to 5.5, similar to the pH increase noted in experimental studies (Table 2). Based on the Henderson-Hasselbalch equation, with a 1 unit increase in lysosomal pH, the ratio of unionized to ionized CFZ would increase 10-fold (from 3 nM to 30 nM). And finally, a broken-down lysosomal membrane or lysed macrophage will lead to an equilibrium with the surrounding cytoplasm or extracellular matrix respectively. The pH in the presence of complete lysosomal destabilization was set to 7.4, leading to lysosomal free base CFZ concentrations approaching its solubility (480 nM).

Extracellular concentrations were varied in the cellular model to estimate a variety of physiologically relevant systemic concentrations. CFZ is both lipophilic and heavily protein bound, leading to a high volume of distribution and a low ratio of free unbound drug. As a result, the extracellular free, unbound drug concentration was varied from 0.01 nM to 10 nM, indicating an expected range of soluble CFZ concentrations in the free base form after termination of drug loading.

The single cell model was developed with four compartments: the lysosome, mitochondria, cytosol, and extracellular matrix (Figure 3A). The corresponding concentrations over time were predicted in each of these four compartments under fixed lysosomal and extracellular concentrations (Figure 3B). The duration of drug activation was estimated based on the predicted flux from the lysosomal compartment, and the previously reported average mass of CFZ loaded into the cell (100 fmols/cell) [7]. After the CLDI had fully solubilized, the respective concentration in the cytoplasm was estimated (Figure 3C). The corresponding time to complete destabilization of drug from the lysosome was then calculated (Figure 3D).

Figure 3.

Computational drug elution model. A) Simulations were performed to compare the cellular pharmacokinetics of CFZ in the presence of four fixed extracellular concentrations (0.01, 0.1, 1, and 10 nM), and three fixed lysosomal concentrations of CFZ (3, 30, and 480 nM). B) The predicted drug concentrations in each of the four compartments are displayed with respect to time at a fixed extracellular concentration of 0.1 nM and variable lysosomal pH. C) The peak predicted concentrations within the cytosol are reported under varying lysosomal and extracellular concentrations. D) Time until total dissolution of CLDIs within the lysosomal compartment are reported under each modeled condition. “Inf” was used to illustrate a negative flux, and therefore no solubilization of drug from CLDIs.

With increasing freely soluble CFZ concentration as a function of increasing lysosomal pH, the resulting duration of drug efflux from the lysosome decreases. As expected, drug is retained in the lysosome for a longer duration of time if the intracellular environment is closer to homeostatic equilibrium. With high enough extracellular concentrations, and limited lysosomal destabilization, the direction of the flux favors accumulation in the lysosomal compartment, indicating that drug would continue to enter the lysosome and precipitate out. However, the more disrupted the lysosomal microenvironment becomes, the faster drug destabilizes in CLDI form, leading to higher flux of drug from the lysosomal compartment out into the cytosol, in parallel to an accompanying increase in free drug concentrations. As expected, the cytoplasmic concentrations of CFZ increase with increasing drug efflux from the lysosome. Complete lysosomal membrane rupture results in large short-lived (hours) rise in cytosolic concentration. Minor lysosomal perturbation results in lower cytosolic concentration for a prolonged duration (days).

3.3. Molecular Impact of Insoluble-to-Soluble Phase Transitions

Using SARS-Cov-2 infection as a clinically relevant candidate therapeutic target for CFZ [59, 64-66], we evaluated the feasibility of this pathogen activated CLDI destabilization mechanism as a prophylactic treatment to prevent or curtail infection. Using the previously cited minimum inhibitory concentration (MIC) of CFZ against COVID-19 (0.32 μM), we evaluated whether the concentrations of CFZ that are predicted by the model exceed the minimum inhibitory concentration (MIC) of SARS-Cov-2 under varying conditions. To run these simulations, the CFZ concentrations in cytosol, lysosome, and mitochondria were predicted under various lysosomal pH and fixed extracellular concentrations, focusing on the effects of CLDI destabilization and the drug concentrations needed for the inhibition of viral replication and spreading (Figure 4).

Figure 4.

CFZ anti-infective potential in the presence of different extracellular concentrations (columns) and virus induced lysosomal membrane destabilization conditions affecting pH homeostasis (rows). Cytosolic (C), mitochondrial (M), and lysosomal (L) compartments were evaluated for COVID-19 anti-infective potential. Clear compartments represent predicted CFZ concentrations at least 1.5-fold below the MIC for COVID-19; yellow represents predicted CFZ concentrations within 1.5-fold of the MIC; and green represents predicted CFZ concentrations at least 1.5-fold above the MIC for COVID-19.

When an uninfected cell with an unperturbed lysosomal environment (lysosomal pH = 4.5) is in the presence of low extracellular concentrations of CFZ, there is a low likelihood of intracellular COVID-19 elimination. Following infection with virus, the extent of CLDI activation increases (by a virus-induced increase in lysosomal pH) or by an increase in the local extracellular concentration. Under these conditions, the likelihood of CFZ concentrations exceeding the COVID-19 MIC in various subcellular organelles continues to increase. Upon complete disruption of the lysosome, say, accompanying the death of the macrophage, cellular concentrations of CFZ exceed the MIC independent of the extracellular concentration. This implies that even in the absence of circulating drug, or long after the last dose of drug has been administered, viral induced cell death of a pre-loaded macrophage has the potential to generate CFZ concentrations that suffice to eliminate a local infection.

This stimulus responsive property was then evaluated against many other infectious pathogens. According to the simulation results, CFZ could be efficacious by directly inhibiting the proliferations and transmission of a large variety of pathogens, by interfering with cellular host protein functions that are necessary for pathogenicity, or by directly interfering with macrophage viability by endowing these cells with a cyto-suicidal, self-defense mechanism (Table 3).

Table 3.

Lysosomal pH dependent inhibition of molecular targets by CFZ. Infectious targets and endogenous targets are inhibited by CFZ to varying degrees with lysosomal destabilization [15, 67-81]. At each lysosomal pH the corresponding molecular target is influenced by CFZ (+), uninfluenced (−), or potentially influenced (+/−) based on the model predicted cytosolic CFZ concentrations at 0.1 nM fixed extracellular concentration. Additionally, cellular interference at high fraction unbound (f_u) or high local concentrations was evaluated by assuming a local extracellular concentration of 10 nM. [15, 64, 67-81]

Molecular Target	MIC/EC50 (μM)	Ki/Kd (μM)	Cellular Inhibition at Lysosomal pH			Inhibition at High fu	ref
			pH = 4.5	pH = 5.5	pH = 7.4
Infectious Targets
Tuberculosis H(37)Rv	0.25 to 0.51		−	+/−	+	+	67
Tuberculosis MDR-TB	0.25 to 4.05		−	+/−	+	+	67
Non-Tuberculosis Mycobacterum (RGM)	2.28 (<0.066 to 8.44)		+/−	+/−	+/−	+	68,69
Non-Tuberculosis Mycobacterum (SGM)	1.16 (<0.066 to >16.88)		+/−	+/−	+/−	+	68, 69
SARS-CoV-2	0.31		−	+	+	+	64
MERS-CoV	1.48		−	−	+	+	64
Gram Positive Bacterium	1.05 to >67.52		−	−	+/−	+/−	70
Gram Negative Bacterium	> 67.52		−	−	−	−	70
Cryptosporidiosis	0.015		+	+	+	+	71
HepG2 CC50	26.43		−	−	−	−	72
Giardia Lamblia	1.8		−	−	+	+	72
Endogenous Targets
Kv 1.3 Potassium Channel Inhibition	0.3		−	+	+	+	73
Caspase-3 Activation	21.1#		−	−	−	+	74
PPAR-γ Activation	0.1		+	+	+	+	75
pRABV glycoprotein G	1.7	4.319	−	−	+	+	76
Cathepsin L Inhibition		71	−	−	−	−	77
CYP3A4/5 Inhibition		0.000786	+	+	+	+	78
CYP2C8 Inhibition		0.00372	+	+	+	+	78
CYP2D6 Inhibition		0.00246	+	+	+	+	78
hRKIP Activation		106*	−	−	+/−	+/−	79
Cell Death
Macrophage (RAW 264.7)	10		−	−	−	+	15
Multiple Myeloma	9.8 (+/− 0.7)		−	−	−	+	80
Chronic Lymphocytic Leukemia	1.0^		−	−	+	+	81

^*Activity shown at 3.2 μM

^#Apoptosis was observed at 21.1 μM concentration, however, the minimal concentration to induce apoptosis may be lower

^{^}In combination with doxazosin, igmesine, or a B-RAF kinase inhibitor, the addition of CFZ initiated apoptosis

In the clinic, CFZ is now being tested as a treatment for several infectious diseases including multidrug resistant tuberculosis and nontuberculous mycobacterial infections. CLDI activation could also allow local concentrations of CFZ to supersaturate and increase beyond the MIC of most tuberculosis and non-tuberculosis mycobacterium, SARS-CoV-2, MERS-CoV, giardia lamblia, and Rabies virus. CFZ also exhibits good in vitro activity against most Gram-positive bacteria species with minimum inhibitory concentrations (MICs) in range 0.5-2 mg/L [82], although Gram-negative bacteria are uniformly resistant to CFZ [64, 82, 83].

CFZ sequestered as CLDIs, in healthy lysosomal microenvironments, is unable to induce apoptosis in macrophage cells [15]. However, high concentrations of CFZ in the free base form, can activate caspase-3, leading to apoptosis [74]. As a result, with increasing CLDI activation, an increased likelihood of cell death is expected. Nevertheless, a CFZ induced (or pathogen induced) cytotoxic chain reaction is mostly expected in granulomas where macrophages are clustered together and loaded with CLDIs and infectious microorganisms, which could translate to improved likelihood of pathogen dependent anti-infective activity.

Indeed, site specific CFZ accumulation has shown to be most pronounced in granulomas [43]. As granulomas can form at sites of infection, a significant bystander effect is expected at these locations. Activation of the mechanopharmaceutical device in one macrophage can facilitate activation in the surrounding cells through local increases in extracellular concentrations. Based on our VCell simulations, granulomas could support a feed-forward, chain reaction mechanism whereby the released CFZ from one cell would likely destabilize the neighboring CLDI-loaded macrophages, interfering with the lysosomal ion homeostasis and thus giving rise to high local concentrations of free drug (Figure 5) [43]. In mice, ablating macrophages with liposomal clodronate inhibits CLDI formation and does not lead to any overt toxicity [43]. Nevertheless, when mice are loaded with CFZ prior to liposomal clodronate injection, CLDIs can be activated by liposomal clodronate, leading to a highly toxic reaction.

Figure 5.

Disruption of lysosomal acidification machinery can lead to different levels of CLDI activation and anti-infective response. Model predicted CFZ concentrations in the lysosome and cytoplasm are reported alongside the hypothesized mechanisms of CLDI destabilization. A) Oral administration of CFZ achieves stable CLDI loading in a healthy cell. CFZ and CFZH⁺ are initially in equilibrium and lysosomal pH is 4.5. B) Local CLDI activation in lysosomes (following an increase in pH to 5.5) inhibits targets mediating pathogen replication, assembly, and egress within the cell. C) Increasing lysosomal pH to 7.4 in the presence of low to moderate CFZ loads leads to activation that suffices to inhibit cellular proteins involved in pathogen replication, assembly, and egress. D) In the presence of a large CLDI load, complete activation induces cell death and then acts to inhibit viral and host mechanisms of neighboring cells.

In a healthy pre-loaded cell (Figure 5A), both CFZ and CFZH⁺ are initially in equilibrium at the normal lysosomal pH of 4.5. Local activation of the lysosome results in increased solubilization of CLDIs, higher concentrations of free base, and faster flux across the lysosomal membrane (Figure 5B). This increase in soluble CFZ leads to the inhibition of targets mediating pathogen replication, assembly, and egress. As a result, an increase in pH brings about localized anti-infective activity. Upon further activation, through membrane penetration, or complete disruption of lysosomal acidification machinery, local freely soluble CFZ concentrations continue to increase (Figure 5C). Under complete destabilization, the lysosomal pH approximates 7.4, matching the neutral internal pH of the cytoplasm. As the lysosomal membrane breaks down, complete disruption of the conditions required for CLDI stability would lead to a bolus of soluble CFZ spreading to neighboring cells (Figure 5D).

3.4. Pre-Exposure Prophylaxis in at Risk Populations

With the stimulus responsive activation of CFZ, a new therapeutic strategy for pandemic prevention could be feasible and warrants exploration. Self-assembling drug depots with targeted and powerful anti-infective potential could allow for the development of broad-spectrum, pre-exposure prophylaxis regimens. The concept of pre-exposure prophylaxis for infectious pathogens has been adopted for HIV [84, 85] as well as considered for SARS-CoV-2 [86, 87]. While pre-exposure prophylaxis has been shown to be efficacious, patients are also exposed to relatively high levels of circulating drug, making long-term usage of prophylactic regimens potentially hazardous. As evident in the low toxicity of CFZ, the biocompatibility of CLDIs makes cytoplasmic construction of mechanopharmaceutical devices a clinically feasible prophylactic option. Accordingly, testing cytoplasmic construction as a prophylactic approach requires a completely different strategy from the short, post-infection treatment regimens that are being tested in clinical trials, as was recently the case with cryptosporidiosis [88]. If CFZ is to form a pathogen-activatable intracellular drug depot, pre-loading healthy individuals with the drug prior to infection is essential for the drug to accumulate in the macrophages and self-assemble into CLDIs. As an emerging field of study, careful experimental design and in-depth mechanistic analysis would be essential for understanding conditions governing the trigger mechanisms leading activatable drug release [89].

4. Properties of Self-Assembling Molecules

While CFZ is an inspiring drug, it cannot be expected to be a panacea to protect humanity against all pathogens. For this stimulus responsive self-assembling treatment approach to become, and remain viable, additional molecular entities should provide alternative starting points for mechanopharmaceutical device design and lead development efforts. Although the vast majority of small molecule drugs in clinical use are weak bases, CFZ is the only weakly basic drug which is known to accumulate and self-assemble as hydrochloride biocrystals in macrophages of mice and humans. Although lysosomotropic property of CFZ is common to weak base compounds, the extent of drug accumulation in lysosomes could vary significantly even for molecules with similar properties, such as pK_a, logP, and intrinsic free base solubility [90]. For many drugs, lysosomal accumulation is associated with toxic effects, such as perturbation of lysosomal ion homeostasis. These effects include increasing lysosomal pH, increasing the lysosomal size, and permeabilizing the lysosomal membrane [82, 90-93]. In the case of hydrophobic weak bases like amiodarone, carvedilol, and desipramine [94-96], the estimated degree of supersaturation of the salt form of the drugs in the lysosomal environment would be 1 to 12 orders of magnitudes lower than that of CFZH⁺Cl⁻, which would decrease the propensity of these other drugs to precipitate out and self-assemble.

Thus, amongst FDA-approved weakly basic drugs [94-96], CFZ is unique because of its extremely low solubility in the lysosomal microenvironment. Nevertheless, there may well be many other different in vivo self-assembling molecular entities that could serve as drug candidates and are awaiting discovery as molecular building blocks for cytoplasmic construction. In terms of alternative mechanisms that do not necessarily involve a lysosomal ion trapping and precipitation pathway, the anomalous pharmacokinetics of anacetrapib exhibits the kind of behavior that would be expected from a soluble-to-insoluble phase transition phenomenon.

5. Conclusions

There are many cognate research areas that are relevant to cytoplasmic construction. For example, it may be feasible to identify functional supramolecular structures in high throughput screening experiments of biased libraries of small molecule chemical agents that are selected based on their putative ability to undergo soluble-to-insoluble phase transitions in the microenvironmental conditions of specific organelles. Alternatively, the de novo computational design of functional supramolecular structures that can self-assemble from certain classes of building blocks may yield promising drug candidates. Additionally, new computational modeling approaches within the realm of pharmacokinetics, cheminformatics, crystal engineering, and molecular dynamics simulations could all be relevant to advance cytoplasmic construction as a scientific and bioengineering field. Furthermore, biointerface science, including the design of specific proteins that bind to and interact with an intracellular mechanopharmaceutical device to transmit mechanical, optical, electrical, and other signals between the mechanopharmaceutical device and various cellular components is yet another relevant research area that has yet to be explored. Thus, cytoplasmic construction offers plenty of research opportunities in terms of exploring the self-assembly and interactions of intracellular mechanopharmaceutical devices with the natural components of cells.

The extraordinary transport properties of CFZ already allow this broad-spectrum anti-infective agent to accumulate as inert precipitates within macrophage lysosomes in a manner that warrants further clinical testing as the first, broad spectrum biodefense approach to pandemic protection. Certainly, the stimulus responsive nature of CLDIs lends itself to the exploration of a new paradigm for anti-infective drug development. Self-assembling mechanopharmaceutical devices could be used to confer pre-exposure prophylaxis in the setting of high-risk transmission of pandemic level pathogens, such as hospitals or nursing homes. CLDIs and other mechanopharmaceutical devices are certainly worthy of further exploration as stimulus responsive intracellular drug depots.