Formulations for Pulmonary Administration of Anticancer Agents to Treat Lung Malignancies

Methotrexate

Although not used alone for the primary treatment of lung cancer, methotrexate has been used as proof of concept for the inhalation delivery of chemotherapeutic agents using a pMDI delivery system.⁽⁵⁸⁾ Chlorofluorcarbon (CFC)-free formulations were prepared by cryomilling methotrexate, with or without Poloxamer 217 as a physical mixture. The low solubility of this drug in 10% ethanol/hydrofluoroalkane (HFA) 134a gave rise to a suspension dosage form. As investigated by X-ray diffraction, the grinding process did not alter the crystalline drug morphology. The first evident result from this study is that cryomilling only was not an effective method to reduce particle size for pulmonary delivery of methotrexate. The highest respirable fraction (RF) was achieved by taking the cryomilled particles and further processing by sieving (5-μm sieve). The values were as high as 35% RF for this formulation, whereas nonsieved preparations did not greatly vary from about 15% RF, regardless of grinding media and duration of process. Respirable fractions were defined as the percentage of particles deposited in stages of an Andersen cascade impactor with diameter less than 4.7 μm, to the total amount of drug emitted from the device (including all impactor stages, throat, and actuator). These results show that characterization of particle size distribution prior to formulation is highly beneficial to assess the potential to deliver the drug to the lungs. In general, the formulations presented MMADs of 2.2 to 3.2 μm and GSDs >2. Undoubtedly, a greater reduction in particle size would be necessary to decrease GSD and consequently improve the respirable fraction results. Nevertheless, with this model anticancer drug, the investigators found in vitro cytotoxicity results of greater than 50% cell kill and comparable results between aerosolized and nonaerosolized doses were obtained. This outcome was sufficient to encourage further investigation of the pulmonary route for the delivery of other chemotherapeutic agents.

COX-2 inhibitors

The aerosol characteristics of different combinations of anticancer agents and selective COX-2 inhibitors against NSCLC cell lines have been examined for the in vitro cytotoxicity. For these studies, aerosolized nimesulide and aerosolized celecoxib have been formulated to evaluate cytotoxicity potentiation in conjunction with, respectively, doxorubicin and docetaxel.^(59,60) The COX-2 inhibitors nimesulide and celecoxib are practically insoluble in water and slightly soluble in ethanol.⁽⁶¹⁾ After a solubility study in both HFAs 134a and 227 of the drug to be aerosolized, solution formulations for pMDI delivery were prepared. Water-soluble drugs can be simply dissolved in an aqueous mixture prior to nebulization. Alternatively, solutions may be prepared for pMDI devices, by dissolving the drug in the propellant. In these cases, given the relatively recent change of propellant type from CFC to HFA,⁽⁶²⁾ it may be necessary to have a cosolvent (to dissolve the drug) that is miscible with HFA. Ethanol can often fulfill these two requirements. Stages 3 to 6 of the cascade impactor (particles with cutoff diameters of less than 4.7 μm) were used to assess respirable fraction. The formulations, the devices, and the particle size characterization are shown in Table 2.

The formulations of COX-2 inhibitors presented in Table 2 demonstrated feasible MMADs for drug delivery to the lungs, although with a relatively broad GSD for nimesulide formulation. From the studies with Celecoxib, the formulations with higher drug loading have presented higher medication delivery. Also, it has been previously shown that as ethanol concentration decreases, either in HFA 134a or 227, the fine particle fraction increases.⁽⁶³⁾ This event has also been observed for aerosolized celecoxib formulations. Importantly, the aerosolization process has been reported not to alter the activity of the aerosolized drug. In addition, these two formulations presented satisfactory physical and chemical stability for 1 month at room temperature, and elevated temperature (40°C). The celecoxib formulation also presented a similar stability at room temperature only. The aerosolized selective COX-2 inhibitors have been shown to satisfactorily decrease IC₅₀ values of the aforementioned anticancer agents in lung cancer cell lines. But, are COX-2 inhibitors potent enough to be delivered via pMDI? In the case of COX-2 inhibitors, each actuation was able to deliver about 50 to 100 μg of drug. From this, only about 40 to 50% was potentially able to reach the deep lungs. That means that at least 20 actuations may be needed to achieve about 1 mg of COX-2 inhibitors to the lungs. Once more, we emphasize the dose restriction of pMDI delivery of anticancer drugs. Therefore, the capability of pMDI formulations of COX-2 inhibitors to effectively treat lung cancer in conjunction with other chemotherapeutic agents is yet to be answered.

Celecoxib has also been prepared as an emulsion for inhalation delivery via nebulizer.⁽⁶⁴⁾ Based on a previous study,⁽⁶⁵⁾ the antitumor activity was tested in a human orthotopic NSCLC xenograft model. Aerosolized celecoxib was compared with oral delivery, both concurrently with an intravenous administration of docetaxel. Celecoxib (5 mg/mL) was dissolved in ethanol and polyethylene glycol (PEG) 400 before emulsification with molten D-α-tocopheryl polyethylene glycol 1000 succinate (vitamin E TPGS). Vitamin E TPGS has previously demonstrated to increase drug bioavailability via oral administration by inhibition of the multidrug resistant transporter P-glycoprotein.⁽⁶⁶⁾ The expression of P-glycoprotein in NSCLC patients is activated during chemotherapy.⁽⁶⁷⁾ Therefore, the use of vitamin E TPGS could potentially be more important than simply acting as an emulsifier agent. On the other hand, to our knowledge, there has been no report so far on the safety of pulmonary administration of this excipient. A Pari LC Star^® jet nebulizer was used to generate particles with MMAD of 1.68 μm and GSD of 1.36 as measured by a Mercer Cascade Impactor (airflow rate not reported). Female Nu/Nu mice were exposed to aerosolized drug using a nose-only inhalation chamber for 30 min per day during 28 days. According to the authors, the estimated total deposited dose was 4.56 mg/kg/day; this calculation was based on concentration of drug in aerosol volume, volume of air inspired by the animal during 1 min, estimated deposition index for mice and duration of treatment. Compared to a control group, reduction in the tumor volume was 61 and 54%, following aerosolized and oral treatments in combination with i.v. docetaxel, respectively. The accuracy of the estimated deposited dose is questionable. To explain, this dispersion system may be considered to present similar thermodynamic characteristics to that of liposomes. Due to the stress applied, liposome structure is significantly disrupted by jet nebulizers.⁽⁶⁸⁾ In the study of aerosolized celecoxib emulsion, the system behavior under aerosolization via jet nebulizers was not evaluated. This could raise questions as whether the drug was homogeneously released from the device throughout the duration of exposure. Nevertheless, other factors are very relevant from the formulation aspect and therefore should also have been evaluated, such as the particle size distribution of the emulsion and its stability over time.

Gemcitabine

In combination with other antineoplastic agents, gemcitabine has shown effectiveness against a number of types of cancer, including advanced stages of NSCLC.⁽⁶⁹⁾ A salt form is currently commercialized under the trade name Gemzar^® (Gemcitabine HCl) by Eli Lilly (Indianapolis, IN). Gemzar^® is supplied as a freeze-dried powder with mannitol and sodium acetate, hydrochloric acid, and sodium hydroxide as excipients. Reconstitution with saline gives the final solution of this marketed injection formulation, with pH ranging from 2.7 to 3.3. Clinical and toxicological studies have been performed by drug aerosolization using this solution for injection in mice, rats, dogs, and baboons.^(70–74) The most important features of these studies can be seen in Table 3.

The accuracy in the characterization of aerodynamic profile and drug deposition in some of these studies are questionable, particularly as one of the studies fails to mention the specified type of nebulizer used. Nonetheless, these gemcitabine preclinical studies give very significant therapeutical and toxicological data about inhaled solution for both primary and metastatic lung cancer. For instance, five weekly treatments of 8 mg/kg to mice bearing human orthotopic NSCLC tumors were able to significantly inhibit the primary cancer growth. Meanwhile, increasing the dose to 12 mg/kg caused acute and fatal pulmonary edema events.⁽⁷¹⁾ Similarly, lung metastases incidence was significantly reduced by doses of 1.0 mg/kg biweekly for 6.5 weeks.⁽⁷⁰⁾ Also, gemcitabine has been shown to be safe in rats and dogs at a MTD of 9 weekly treatments of 4 mg/kg and 20 biweekly treatments of 2.38 mg/kg, respectively.^(72,74) Pharmacokinetic studies in baboons have shown that t_max is achieved at approximately 10 min after starting nebulization (in this case, still during inhalation process).⁽⁷³⁾ After 60 min, the drug was undetectable in the plasma; not a surprising result, given the low partition coefficient of gemcitabine (log P = −1.24).⁽⁷⁵⁾

Despite the number of studies in different species, so far, only the commercialized injection formulation of this anticancer agent has been investigated for the possibility of aerosolization to the lungs. With an acidic pH, the desired water dissolution of gemcitabine for injection (pK_a = 3.58) is assured.⁽⁷⁶⁾ When inhaled, the prompt drug availability for lung absorption will highly contribute to a rapid pulmonary clearance by diffusion mechanism. On one hand, the hydrophilic characteristic of gemcitabine provides an advantageous lung residence time compared to its lipophilic chemotherapeutic counterpart. On the other hand, more investigation exploring its water solubility properties are yet to be performed so as to ascertain the full potential of gemcitabine to treat lung malignancies.

Doxorubicin

Doxorubicin is a relatively lipophilic and water-soluble drug that has long been used for the treatment of pulmonary malignancies. Based on its physicochemical characteristics, namely, water solubility and partition coefficient (see Table 1), one would expect a prompt absorption of doxorubicin by the lungs.

A DPI formulation was studied to evaluate the in vitro cytotoxicity of doxorubicin nanoparticles.⁽⁷⁷⁾ The drug was incorporated into poly(butylcyanoacrylate) nanoparticles by emulsion polymerization and coated with polysorbate 80 (0.5 % v/v). Poly(butylcyanoacrylate) had previously been used for the intravenous administration of anticancer drugs for the treatment of brain tumors.⁽⁷⁸⁾ Inclusion of a surfactant coating has been shown to facilitate particles translocation across the capillary barrier.⁽⁷⁹⁾ The doxorubicin-loaded nanoparticles were then mixed with lactose (carrier) in a spray freeze-drying procedure. This manufacturing process, which may be scaled up, was able to yield a loading capacity of 1.39 μg of doxorubicin per mg of powder. The results from this study showed that doxorubicin-loaded nanoparticles of 173 ± 43 nm mean particle sizes were effectively incorporated into a 10 ± 4 μm overall lactose geometric particle size. Using a passive DPI, the MMAD obtained was 3.41 ± 0.22 μm, as measured at 60 L/min by Mark II Anderson Cascade Impactor. This powder formulation has demonstrated an increased cytotoxicity against A549 and, more significantly, against H460 lung cancer cell lines compared to free drug. However, the presence of polysorbate 80 in the formulation does not provide confirmation that endocytosis of doxorubicin-loaded nanoparticles has occurred more readily due to drug particle size alone or by some improvement in the translocation of these particles across the cancer cell barrier. Interestingly, the blank formulation (drug-free nanoparticles) showed certain levels of cytotoxicity as well. Although similar low levels of polysorbate 80 has been safely aerosolized to mice,⁽⁵⁶⁾ to our knowledge, there has been no report to date verifying the safety of delivering poly(butylcyanoacrylate) to the lungs. Therefore, despite the encouraging cytotoxicity and aerodynamic results as well as the possibility of process scale up, this formulation may find restrictions for pulmonary delivery. Nevertheless, more studies are yet to be performed in order to define the feasibility of nanoparticle uptake by cancer cell lines.⁽⁴²⁾

Dogs with naturally occurring primary or metastatic lung tumors were also treated with doxorubicin.⁽⁸⁰⁾ A solution was prepared by dissolving the drug in 20% ethanol at a concentration of 16 mg/mL. Under anesthesia, the dogs were exposed to the drug during a certain period of time by means of a Pari LC^® jet nebulizer coupled to an endotracheal tube; this special assembly was designed to prevent fugitive aerosolization. The animals received six treatments of 3 mg of doxorubicin once every 2 weeks. The dose calculation was based, not only on stabilized minute-volume of respiration under light anesthesia, but also assuming a body surface area of 1 m² (as a result small dogs were naturally exposed to a higher drug concentration in this study). Although 22% of the treated dogs (n = 18) achieved partial responses (a tumor volume reduction >50%), no formulation characterization was reported. This lack of information does not allow us to evaluate whether the outcomes were maximized due to formulation and device performances. For instance, airflow rates, viscosity, and surface tension may all influence the aerosol output from jet nebulizers.⁽⁸¹⁾ Nevertheless, no systemic toxicity was evident, while pulmonary toxicity was observed (mainly as a cough).

The encouraging outcomes from the previous study have set the stage for a clinical trial. Doxorubicin was selected to be nebulized in patients with primary or metastatic tumors to the lungs for a phase I study.⁽⁸²⁾ The inhalation solution was provided at 16 and 24 mg/mL in ethanol:water 1:4 (pH 3). The aerosolization system consisted of a Pari LC Plus^® jet nebulizer in a sealed, mouth-only inhalation apparatus (OncoMyst^® model CDD-2a). As a precaution in this study, a demistifier tent with a HEPA filter was setup to protect the health care provider from potential fugitive aerosol. The dose was calculated on the bases of inhaled Technetium 99m deposition and scintigraphy, and a limited pharmacokinetic evaluation was performed in this study. Due to the required number of breaths in the aerosolization system, administration of higher drug concentrations resulted in dosing times as long as 45–60 min. During this period, no drug plasma levels were analyzed. Maximal plasma concentration was observed at the first sampling point, 5 min after finishing treatment. Similarly to the study in dogs, no systemic toxicity was observed, whereas cough was the most frequent adverse event observed in the patient population. Dose-limiting pulmonary toxicity was observed at 7.5 mg/m² every 3 weeks. Therefore, dose escalation of this formulation by pulmonary administration has been suggested not to be adequate. Despite the nominal nebulizer capability to generate aerosol particles of 2–3 μm mentioned by the authors, again, the aerodynamic characteristics of the formulation were not confirmed by in vitro analytical tests. On one hand, the acidic solution (pH 3) present ionizes the doxorubicin (pK_a = 8.34) increasing the drug solubility.⁽⁸³⁾ On the other hand, the drug lipophilicity favors diffusion across membranes, considering the partition coefficient of doxorubicin (log P = 0.65). As a result, the fraction of unionized drug would be quickly absorbed and a relatively low lung residence time would be expected. Therefore, the long dosing time required to administer high drug concentrations may have biased these pharmacokinetic data. Another important aspect from this formulation is its pH. Previous studies have demonstrated that acidic aerosols stimulate coughing and increase lung resistance, with consequent bronchoconstriction.^(84–86) In this study, pulmonary toxicity has been observed at high delivered doses, in so far as higher volume of liquid formulation was deposited. As the formulation pH increases the drug solubility, the acidic solution evokes airway irritation that may potentially be aggravated by increase in volume of liquid formulation deposited from the aerosol. Therefore, this formulation parameter appears as a confounding factor that may compromise the determination of the MTD of doxorubicin.

Doxorubicin is one of the few drugs to advance from preclinical studies to clinical trials for the pulmonary delivery to treat lung malignancies. Whether its intrinsic dose-limiting toxicity may impose restrictions to its use as a single chemotherapy treatment modality is yet to be determined. In our opinion, formulations containing doxorubicin should be more appropriately designed and characterized prior to defining their intrinsic dose-limiting toxicity, before an adequate assessment of the associated benefits of inhalational therapy. This will ultimately determine the role of this chemotherapeutic option as single modality treatment. Furthermore, a neoadjuvant strategy is yet another option that still needs to be investigated prior to defining the role for pulmonary delivery of doxorubicin in lung cancer treatment.

Farnesol

Farnesol is a lipophilic (log P = 5.31) and very slightly water-soluble drug that induces apoptosis in lung carcinoma cells.^(87–89) An emulsion of this anticancer agent was prepared for nebulization by mixing farnesol (10.5 mg/mL) with polysorbate 80 (0.5 mg/mL) in a 20% v/v ethanol/normal saline mixture.⁽⁹⁰⁾ Emulsion stability was verified by storing a mixture of farnesol, ethanol, and polysorbate 80 at 4°C for up to 14 days, with saline added just before use. The aerodynamic profile of this formulation was measured by a Phase Doppler Anemometer coupled to a breath simulator (average airflow rate of 18 L/min and tidal volume of 0.75 L).⁽⁹¹⁾ According to this method, Pari LC Star^® and LC Plus^® nebulizers were able to generate aerosol particles with MMADs of 4.96 μm and 6.87 μm and GSDs of 1.48 and 1.67, for each nebulizer, respectively. Also, the nebulization efficiency of this emulsion was estimated to be approximately 30% deposited drug out of the total volume submitted to aerosolization for either device. Finally, in vitro cytotoxic effects of nebulized farnesol were observed against A549 and H460 lung cancer cells. In our opinion, these data suggest that a low drug deposition in vivo as well as fast lung clearance may be expected. First, only one-third of drug amount from the emulsion volume was effectively aerosolized by these jet nebulizers. Considering solution dosage forms, both jet and ultrasonic nebulizers have similar drug delivery performance to the lungs. When it comes to suspensions though, drug aerosolization is more effective using jet nebulizers because the ultrasonic effect may aerosolize empty droplets (no drug present).⁽⁹²⁾ Now, fragile disperse systems like emulsions may be vastly disrupted by the high energy imparted by jet nebulizers. Therefore, the amount of drug that ends up reaching the deep lungs may be compromised. Second, the aerodynamic results of this system are not very favorable for deposition in the lower airways. With MMAD values varying from 5 to 7 μm, most of the aerosol droplets would be deposited in the throat and upper airways, subsequently being swallowed. Finally, farnesol is a very lipophilic drug, with a log P value considerably higher than zero; an almost instantaneous absorption is more likely to occur.

In case of poorly water-soluble drugs, emulsification may be a preferred strategy for delivering the agent to the lungs. However, some important formulation characteristics should always be evaluated. The particle size distribution of the emulsion internal phase may be determined by the energy input during nebulization process. The emulsion particle size may greatly influence the final droplet size aerosolized. Also, it is important to determine whether the type of device chosen for aerosolization will be able to gently deliver the drug, without disrupting the emulsion formulation. In addition, the excipients and production method applied may influence the short- and long-term emulsion formulation stability. In selecting the excipients, the scientist should strive to extend the drug release from the delivery system while being aware of possible toxic effects due to low clearance rates. Considering its physicochemical properties, the potential of farnesol as an anticancer agent to treat lung malignancies via inhalation may be improved with different formulation designs.

Paclitaxel

In the study of primary or metastatic lung tumors naturally occurring in dogs being treated with doxorubicin, a paclitaxel solution was also investigated.⁽⁸⁰⁾ The taxol (75 mg/mL) was dissolved in a mixture of polyethylene glycol 200 and ethyl alcohol. Using the same drug deposition system and regimen, the doses ranging from 10 to 90 mg/m² were administered with no signs of systemic toxicity. As opposed to doxorubicin, local toxicity was not evidenced either. Out of 15 treated dogs, only one partial and one complete response were observed. The drug plasma levels after aerosolization for both doxorubicin and paclitaxel were less than one-tenth of plasma levels observed in normal dogs after intravenous injection of therapeutic range doses. However, the lack of information about the formulation characteristics makes it difficult task to speculate on the performance of this solution.

Drugs formulated within biodegradable polymers can present pharmacokinetic profiles with a delayed and/or extended release compared to the free drug. This sustained-release system can be controlled based on size and porosity of prepared spheres.⁽⁹³⁾ The feasibility of delivering anticancer agents within this type of dosage form has been investigated using paclitaxel.⁽⁹⁴⁾ Microspheres of the polymer poly-(L-glutamic acid) (PGA) were loaded with 20% (w/w) paclitaxel (PGA-PTX). For comparison purposes, 0.6% w/w paclitaxel was dissolved in cremophor, making a similar taxol solution to the commercially available product (Taxol^®, Bristol-Myers Squibb). Aerosol output ratio tests demonstrated that PGA-PTX formulation was able to deliver 80–400 times faster than taxol solution. Although the drug was only deposited in mice lungs via intratracheal injection, the aerosol characteristics was evaluated when generated by a Salter Labs^® 8900 jet nebulizer at 5 or 9 L/min. Using a seven-stage cascade impactor, the authors reported that approximately 50% (wt) of the aerosolized polymer formulation presented droplets with less than 5 μm. However, the amount of drug in each stage was measured by weighting the filters pre- and postaerosolization instead of using an analytical method such as an UV spectrophotometer. Notably, this can render significant measurement errors. First, the authors did not consider the differences in concentration from the taxol solution (6 mg/mL) to the PGA-PTX formulation (not specifically mentioned, but ranging from 3 to 25 mg/mL). Second, the instrumental errors generated by the balance to measure formulation amounts in the milligrams were neglected. Therefore, the aerosol characteristics of this formulation are highly questionable. Nevertheless, the authors reported a concentration- and time-dependent in vitro cytotoxicity against H358 and H460 NSCLC cell lines. In addition, after single intratracheal administration, a MTD of 30 mg/kg (paclitaxel equivalents) was observed. Three weekly doses of taxol solution (2.5 mg/kg) or PGA-PTX (20 mg/kg paclitaxel equivalents) were both sufficient to significantly improve survival at about 135 days. Although the results may seem encouraging for further preclinical studies with pulmonary delivery of this formulation, the authors did not consider a potential accumulation of the polymers in the lungs. Due to a 10-fold greater dose of PGA-PTX needed to provide a similar survival as with the taxol solution, the dosage regimen outlines multiple administrations needed. Therefore, the PGA lung accumulation may be significantly high and could consequently compromise even more the pulmonary function of the lung cancer patients.

The pharmacokinetics and the therapeutic efficacy of a liposomal paclitaxel aerosol have also been investigated against pulmonary metastases in a murine renal carcinoma model. Because the excipients used are mainly surfactants naturally occurring in the lung fluid (e.g., phosphatidylcholine), liposomes present a relatively low local toxicity. Also, given their lipophilic nature, formulations of this type are able to encapsulate poorly water soluble drugs.⁽⁹⁵⁾ Liposomes of paclitaxel (10 mg/mL) were composed of dilauroylphosphatidylcholine (DLPC) at drug-to-lipid ratio of 1:10 (w/w). Both components were prepared in t-butanol and subsequently lyophilized.⁽⁹⁶⁾ Prior to use, the formulation was reconstituted with sterile water and vortexed until homogeneously dispersed. With this method to produce liposomes, scaling up is feasible. However, as there was no energy input to form the liposomes (e.g., high-pressure homogenization), it is not surprising that particle size analysis revealed a wide size distribution (2.0 to 25.3 μm). An Aeromist^® jet nebulizer generated aerosols with an MMAD and a GSD of 2.2 μm and 1.9, respectively, as measured using an Andersen Cascade Impactor at 10 L/min. Interestingly, the shear force applied by the jet nebulizer was able to disrupt the liposomes during the nebulization process. The particle size was reduced drastically to less than 0.4 μm. Despite that, the authors estimated the dose to be of 5 mg/kg by chamber aerosol exposure for pharmacokinetic studies. Compared to intravenous injection, the AUC in the aerosol treated group was 26-fold higher, with distribution half life of only 0.71 h. Nevertheless, treatment three times a week over 2 weeks with paclitaxel-DLPC aerosolized liposomes, resulted in significant reduction in tumor number and an increased survival time compared to untreated and DLPC-only-treated mice. These results were encouraging for further studies.

Paclitaxel-induced resistance is related to over expression of plasma membrane glycoprotein (P-gp), which acts as an efflux pump decreasing the intracellular drug concentration. Cyclosporin A (CsA) has the capacity of reversing this resistance, and, when coadministrated with paclitaxel, CsA also acts as an inhibitor of cytochrome P450-mediated metabolism (enzymes that are involved in the metabolism of this anticancer agent).⁽⁹⁷⁾ These findings and a dose-limiting toxicity not mentioned in the previous study prompted evaluation of the antitumor effect of coadministration of CsA and paclitaxel in the same animal model. Paclitaxel liposomes were prepared as mentioned above, whereas CsA liposomes differed only by the drug to lipid ratio (1:7.5 w/w) and concentration (5 mg/mL). Similar to the aerosol characteristics measurement in the study above, CsA-DLPC liposomes presented an MMAD of 1.6 μm and a GSD of 2.2. Chamber aerosol exposure provided an estimated dosage deposition of paclitaxel and CsA liposomes to the lungs of 7.8 and 6.1 mg/kg, respectively. CsA/paclitaxel-treated animals showed lung weights and tumor surface areas significantly lower than paclitaxel-only and untreated groups. Dose escalation of CsA has shown to be more effective to decrease area and number of tumors. However, the treatment (three times per week for 2 weeks) had to be discontinued during toxicity studies due to expressive systemic toxicity (body weight loss).⁽⁹⁸⁾

In terms of safety, the use of phospholipids as excipients to aid the drug delivery to the lungs is a viable option. In addition, they have shown to improve particle migration to the lung periphery owing to the reduction in surface tension provided by these surfactants.⁽⁴⁰⁾ In theory, this applies especially to poorly water-soluble drugs, such as the taxols. However, the encapsulation efficiency of any method to produce liposomes is an issue that should always be addressed. Previous studies have shown that, using the same process with similar phospholipids, somewhat low encapsulation efficiencies were achieved at a constant drug concentration (<45%).⁽⁹⁹⁾ When seeking for a controlled release, not only the encapsulation efficiency becomes important, but the selection of phospholipids may be crucial. Transition temperatures of phospholipids are highly dependent on the saturation and extension of the carbon chain.⁽¹⁰⁰⁾ For instance, DLPC (12 saturated carbons) presents a transition temperature of about −1°C. At the temperature of the human body, the fluid nature of this surfactant will promptly release the drug. On the other hand, dipalmitoylphosphatidylcholine (DPPC-16 saturated carbons), with a transition temperature of approximately 41°C, may control the drug release for the encapsulated drug. In the study above, the lipophilic nature of paclitaxel together with the fluid-like characteristics of DLPC at body temperature justifies the low lung residence time.

After it has been shown in vitro that paclitaxel efficacy is more related to exposure time than to increased dose, a conjugation of paclitaxel with fullerene (C₆₀) was developed to sustain the release of the anticancer drug at the site of administration. Liposomes of paclitaxel-C₆₀ conjugates have been formulated with DLPC similarly to the previous method. A mean diameter of 2.77 μm as measured by light scattering was determined for the fullerene-DLPC-paclitaxel liposomes. Cytotoxicity tests against A549 lung cancer cell lines have demonstrated IC₅₀ values similar to those of DLPC-paclitaxel liposomes.⁽¹⁰¹⁾ Impressively, aerosolization of this excipient to the lungs 3 h per day for 10 consecutive days has presented minimal toxicity. Even so, lung half-lives for nano- and microparticles of C₆₀ were 26 and 29 days, respectively.⁽¹⁰²⁾ These results hold promise of increasing exposure time of the anticancer agent at a tumor site, although safety studies are warranted. Pharmacokinetic and antitumor activity studies of C₆₀-paclitaxel are still being performed in vivo by the same group.

Conversely to previous studies of COX-2 inhibitors where docetaxel was studied along with aerosolized celecoxib, aerosolized docetaxel in conjunction with celecoxib has also been evaluated.⁽¹⁰³⁾ Considering the similarities in their physicochemical properties and formulation design, aerosolized docetaxel study will not be further discussed. The formulation and its characteristics are summarized in Table 2.

Camptothecins

Camptothecins are topoisomerase I inhibitor drugs that demonstrate significant toxicity, especially myelosuppression.⁽¹⁰⁴⁾ Camptothecin (CPT) and its likewise poorly water-soluble derivative, 9-nitrocamptothecin (9NC), also known as rubitecan, were some of the first anticancer drugs to be aerosolized to animal lungs. Similar to liposomal paclitaxel, CPT (10 mg/mL) and 9NC (100 mg/mL) were prepared with DLPC and their anticancer effect tested against human lung cancer xenografts in mice. Following lyophilization and reconstitution, an Aerotech II^® nebulizer operating at 10 L/min aerosolized the formulation into a whole-body exposure chamber. In this study, the formulations were characterized for liposome encapsulation efficiency by Percoll^® gradient analysis. In summary, the studies demonstrated that an increase in drug concentration required a higher proportion of lipids for efficient drug incorporation. Rubitecan concentrations varying from 0.1 to 1.0 mg/mL, in the nebulizer reservoir, resulted in MMADs ranging from 0.8 to 1.6 μm. The encapsulation efficiency of 9NC and CPT liposomes at a drug to lipid ratio of 1:50 (w/w) and 0.5 mg/mL were approximately 80%. To achieve high encapsulation efficiency, the drug concentration was compromised and these formulations were chosen to evaluate their anticancer effect. The MMAD of the 9NC formulation was 1.2 μm. Significantly reduced tumor growth was shown in animals receiving an estimated 9NC dosage of 76.7 μg/kg/day, 5 days per week for 35 days. CPT-treated animals also presented decrease in tumor volume, although 9NC was more effective. Similarly to the liposomal paclitaxel formulation, a shear effect promoted a higher than sixfold decrease in mean size of liposome after nebulization process. Nevertheless, this study also demonstrated that aerosol exposure of 9NC liposomes were more effective to reduce tumor growth than by oral administration. This is a significant indication that the drug absorbed following oral administration, due to mice grooming has very little antitumor effect on lung carcinomas.⁽¹⁰⁵⁾ Finally, evaluation of the effect of delivery route using nose-only aerosol exposure or intramuscular administration indicated that the latter showed a significant decrease in tumor volume compared to untreated animals; however, this was not as nearly effective as with nose-only aerosol exposure. Delivering the drug directly to the lungs has therefore been shown to be essential for the effective reduction in the tumor growth. No myelosuppression signs were observed in this study, and 9NC liposomes have also been shown to be effective against murine melanoma and human osteosarcoma pulmonary metastases in mice.⁽¹⁰⁶⁾

Following this study, the authors reported the pharmacokinetics of CPT in mice after liposomal aerosol inhalation. The formulation containing CPT (0.5 mg/mL) to DLPC ratio of 1:50 (w/w) presented an MMAD of 1.6 μm and a GSD of 2.1, as measured using an Andersen Cascade Impactor. The same device model used previously aerosolized an estimated dose of 80.9 μg/kg after 30 min of nebulization. CPT concentrations in the lungs after aerosolization using either nose-only or whole-body exposure chambers were up to 16-fold greater than that found in the blood. Intramuscular solution injection (233 μg/kg) provided only trace amounts of CPT that could be detected in the lungs, even 2 h after administration when 80% of the drug had been released from the site of injection. The authors compared this aerosolization study with previous reports of camptothecin concentration in the lungs after oral, intravenous, and intramuscular administration. Aerosol treated mice in 50-fold lower concentrations presented 7- to 10-fold higher concentrations in the lungs after 30 min of exposure, compared to the other routes of administration.⁽⁴⁷⁾ As discussed earlier, the doses reported may be the best estimation possible rather than an accurate measurement. Despite this, subacute toxicity of the 9NC liposome aerosol in beagle dogs was not observed after an estimated aerosol dose of 24.7 μg/kg/day, 5 days per week, for 8 weeks.⁽¹⁰⁷⁾ DLPC-only liposomes were determined to be nontoxic as well. Following these results, a phase I study has shown that advanced pulmonary malignancy patients tolerated this treatment well in doses of 13.3 μg/kg/day, 5 days per week for an 8 week period.⁽¹⁰⁸⁾ However, the aerodynamic characteristics of the formulation were not reported in this case. Nevertheless, the doses were administered by a mouth breathing-only face mask with a HEPA-filtered airborne scavenging tent covering the patient and nebulization system, to protect the healthcare professional. Finally, the authors recommended the aforementioned dosage regimen for phase II studies, which have not yet been reported.

Rubitecan liposomes of soybean lecithin and cholesterol were also formulated for different studies: in vitro release, biodistribution in mice, and local toxicity in rats.⁽¹⁰⁹⁾ The thin-film hydration method was used, followed by filtration (0.45 μm) prior to lyophilization. The liposomes presented Z-averages of less than 200 nm as measured by dynamic light scattering. For the in vivo studies, a 9NC solution (0.25 mg/mL) was prepared in dimethyl sulfoxide (DMSO)/PEG400 mixture for comparison purposes. Although the authors failed to report the drug concentration and drug-to-lipid ratio of the liposomes, which limits the discussion herein, entrapment efficiencies greater than 90% were obtained before and after freeze-drying, as measured by centrifugation method. The sustained release properties of the formulation were also verified. Using the dialysis method, the drug release from the liposomes was 32.5% in 1 h and approximately 90% after 24 h. Following intratracheal administration of 0.8 mg/kg, the biodistribution study in mice showed that the mean lung residence time of 9NC liposomes and solution was 1.24 and 0.37 h, respectively. In addition, the sustained release formulation presented an AUC in the lungs 3.4-fold higher than 9NC solution, and 4.73 times when compared to that of an intravenous administration. Finally, intratracheal instillation of rubitecan liposomes demonstrated lower toxicity than the solution dosage form, as investigated by histological studies in rats. These results illustrate that the use of soybean lecithin and cholesterol may be a feasible option to improve lung residence time of anticancer drugs. Yet, aerodynamic characterization of the nebulizer-formulation system followed by further preclinical studies is still needed to evaluate the potential of delivering this formulation to human lungs.⁽¹⁰⁹⁾

Cisplatin

Platinum derivatives, such as cisplatin and carboplatin, are the most traditional drugs used for the treatment of lung malignancies. Sustained-release Lipid Inhalation Targeting (SLIT) cisplatin has been investigated in a phase I study of aerosol dosage form for the treatment of primary and metastatic lung cancers. Liposomes of cisplatin (1 mg/mL) were composed of DPPC and cholesterol at drug to lipid ratios of 1:16 and 1:7.5 (w/w), respectively. Sodium chloride was present as iso-osmotic agent and to enhance the stability of cisplatin. The PARI LC Star^® nebulizer generated aerosols at 15 L/min with an MMAD of 3.7 μm and a GSD of 1.9, as measured using a Next Generation Impactor (NGI) at 5°C and 15 L/min. The impactor refrigeration increases the relative humidity in the inner environment of the NGI to close to 100%, diminishing droplet evaporation prior to impaction.⁽¹¹⁰⁾ In this study, the dose-limiting toxicity (DLT) could not be reached after dose escalation using different strategies: increasing dose level, reducing interval between cycles, increasing number of nebulization sessions per day, and increasing amount of drug inhaled (by flow rate increment of the compressor). Although the liposome size distribution was not reported, the authors evaluated the dispersion stability under nebulization. Approximately 40–50% of total cisplatin is released from the liposomes during the aerosolization process. In this case, the cisplatin liposome structure is being partially disrupted due to the shear force generated by the jet nebulizer, as discussed earlier. Considering the physicochemical properties of cisplatin (water soluble and negative log P), one would expect a relatively slow systemic absorption of the free drug over time. With that, the supposed liposome capacity of slowly releasing the drug may be confounded with the intrinsic characteristics of the antineoplastic agent. In addition, considering that the liposomes are not cleared by macrophages, it is uncertain whether any therapeutic effect may be elicited by the drug that remains encapsulated after deposited in the lungs. On the other hand, the liposomes may possibly be engulfed by the pulmonary macrophages, resulting in no systemic exposure of the anticancer agent, but no therapeutic efficacy either. Nonetheless, stabilization of the disease in 12 out of 17 patients was the best overall response with this treatment.⁽¹¹¹⁾ As with doxorubicin treatment, long administration times were required for inhalational cisplatin therapy, which could reduce patient compliance. This may then impose an alternative criterion, as opposed to local and/or systemic toxicity endpoints. On the other hand, defining a tolerable administration time will certainly restrict high-dose requirements for some antineoplastic drugs. Finally, the authors reported their intention to pursue further studies to define DLT with a higher cisplatin concentration in the liposomes (3 mg/mL) as an alternative strategy. Recently a phase Ib/IIa study was published with relapsed/progressive osteosarcoma metastatic patients. No systemic but only minor local toxicity was observed. The drug was well tolerated by heavily pretreated patients with one year cumulative doses of up to 1020 mg. After the same time period, 2 out of 14 patients remained pulmonary disease free.⁽¹¹²⁾ The DLT of this formulation is yet to be accurately determined. Furthermore, and despite these encouraging results, the effectiveness of this formulation is still to be proven throughout the upcoming phases of the pharmaceutical product development process.

5-Fluoruracil

The neoplastic agent 5-Fluorouracil (5-FU) has long been used for cancer treatment, although not the first line drug to treat lung malignancies. Liposomes, microspheres and Lipid Coated Nanoparticles (LNP) of 5-FU have been prepared to investigate their sustained-release properties, as measured by microdialysis.⁽¹¹³⁾ Lipid-coated nanoparticles consist of a drug-loaded core coated by a lipid shell; the dosage form is dispersed in water in the presence of a surfactant. The submicronized hydrophilic core can contain either the drug alone or in combination with other excipients.

First, DPPC and hydrogenated soy phosphatidylcholine (HSPC) liposomes were prepared by the thin film hydration method, and cholesterol was included to further sustain the drug release. Also evaluated was whether the presence of negatively charged lipids in some formulations, such as dipalmitoyl phosphatidic acid (DPPA) and dipalmitoyl phosphatidylglycerol (DPPG), would promote physical stability by inhibiting liposome aggregation and fusion. Extruded liposomes with vesicle diameter of approximately 0.5 μm presented a higher release constant from a first-order release model than nonextruded liposomes. The presence of DPPA was effective in decreasing the release rate of extruded liposomes, although it was not able to provide the same effect in nonextruded preparations. Among the various liposome formulations studied, drug loading was not higher than 7% wt (measured by centrifugation method) and the drug release was extended to about 8 h. Based on these results and considering the drug lipophilicity, the authors estimate drug administration three or four times daily by inhalation, deeming this formulation ineligible for sustained-release purposes.

In this same study, polymeric microspheres were formulated by spray drying different proportions of the following copolymers: poly-(lactide-co-glycolide) (PLGA), poly-(lactide-co-caprolactone) (PLCL) and poly-(lactide) (PLA). In this study the drug loading was approximately 8% wt and the particle size, characterized by light scattering of the drug in suspension, was 1.2 to 1.5 μm. Based on microdialysis measurements of 5-FU, a release constant was calculated from a first-order release model. The drug released in 24 h was about 70 to 90% of the loaded dose. The results also demonstrated that an increase in the lactide moiety of the polymer progressively decreased the 5-FU release from the microspheres. Microspheres of 5-FU within PLA presented the longest duration of release (>32 h). Importantly, the low 5-FU loading capacity (8%) of this dosage form would require a high amount of polymer to be deposited into the lungs of cancer patients, especially following repeated administrations. The complete degradation time of PLA and PLGA polymers have been reported to be relatively high for this application.⁽¹¹⁴⁾ Therefore, unknown toxicity and respiratory function-related consequences, due to cumulative deposition of these polymers in the lungs, has thus far prevented further development for the delivery of 5-FU microspheres to the lung.

Finally, LNPs of 5-FU were prepared by spray drying the drug with poly-(glutamic acid), poly-lysine, or lactose to form the dosage form core. These cores were further spray dried with various combinations of lipids (tripalmitin, tristearin, cetyl alcohol, and stearyl alcohol) to form an outer shell. The drug, released over 24 h, was about 70 to 90% of the loaded dose for the different formulations studied, with core and total LNP diameters of 500 and 1000 nm, respectively. Based on the release constants, the most appropriate combination of core and shell materials for sustained-release aerosolization formulation was poly-(glutamic acid):5-FU (4:1 wt), and tripalmitin:cetyl alcohol (2:1 wt). The predominance of naturally occurring surfactants (e.g., triglycerides) in this dosage form alleviates the concern about low lung clearance rates of polymers. Thus, although it presented as low drug loading as microspheres (5% wt), the LNPs of 5-FU were chosen to be further studied by the authors. The myelosuppression and bone marrow toxicity caused by 5-FU is expected to be overcome by inhalation delivery of this anticancer agent loaded with LNPs (5-FU LNP). Also, a low release rate of this dosage form would provide a sustained-release delivery system for aerosolized drugs.⁽¹¹³⁾

Following this, investigation of influence of core diameter and lipid shell thickness suggested the latter to be the rate limiting step for the release of 5-FU. Based on this, the authors have developed a release model from polydispersed cores and shells consisting of a sequential zero-order/first-order kinetics. Accordingly, a delivery system consisting of 600 nm diameter poly-(glutamic acid):5-FU (4:1 wt) cores and 200 nm thick tripalmitin:cetyl alcohol (2:1 wt) shells was chosen for in vivo studies in hamsters. The particle size of this formulation measured by dynamic light scattering was 1.02 ± 0.26 μm. From a diluted dispersion, an ultrasonic nebulizer (model not specified) generated aerosol droplets that were subsequently dried to the drug particle unit (reflux drying process). The formulation aerodynamic properties were then measured using an Andersen cascade impactor (airflow not reported). The MMAD was 1.15 μm with a GSD of 2.15 and neither the reflux drying process nor the ultrasonic nebulization demonstrated an influence on the release rate. Hamsters in whole-body aerosol exposure units presented estimated drug deposition in the lungs as high as 3.4 ± 0.3%, based on nebulizer output rate and respiratory minute volume.⁽¹¹⁵⁾

Next, fluorescein isothiocyanate dextran (FITC-dextran) was added to the formulation, and hamsters were exposed to a nose-only aerosol chamber for an eight-component pharmacokinetic modeling study. Separate HPLC analysis of FITC-dextran and 5-FU ensured distinction of entrapped and released 5-FU. LNPs demonstrated a 5-FU peak concentration of 0.13 μg/g at 1.02 h. In addition, the half-life of LNPs in the lungs was 4.95 ± 0.38 h with an almost complete clearance in 24 h. Using pharmacokinetic modeling, the authors estimated that the effective dose (0.065 to 0.13 μg/g, for DNA synthesis inhibition) would be maintained for up to 5.4 h. The pharmacokinetic model used also predicted that, using this route of administration, lung levels would be 5.5 times higher than the systemic circulation.⁽¹¹⁶⁾ In their conclusion, the authors discuss that, when considering greater than 20% drug deposition (which can easily be achieved by modern medical aerosol devices) doses of about 100 mg can be translated to humans.

This thorough 5-fluorouracil formulation study is a great illustration of how the dosage form may potentially improve the drug delivery to the lungs. Selection of excipients combined to careful assessment of drug characteristics and, consequently, well-elaborated formulations are crucial to achieve performances that will indeed explore the full antineoplastic potential of each drug. However, further studies are warranted to confirm these predicted pharmacokinetic modeling data.