Inhaled Anti-Tubercular Therapy: Dry Powder Formulations, Device And

Home , Pretomanid

UNI SYDNEY LOGO

INHALED ANTI-TUBERCULAR THERAPY:

DRY POWDER FORMULATIONS, DEVICE AND

TOXICITY CHALLENGES

Thaigarajan Parumasivam

A thesis submitted in fulfillment of the requirements for the degree of

Doctor of Philosophy

Faculty of Pharmacy

The University of Sydney

2017

STATEMENT OF AUTHENTICITY

This thesis is submitted to the University of Sydney in fulfilment of the requirements for the

Degree of Doctor of Philosophy. The work described was carried out in the Faculty of

Pharmacy and Centenary Institute under the supervision of Professor Hak-Kim Chan and

Professor Warwick Britton

The work presented in this thesis, is to the best of my knowledge and belief, original except as acknowledged in the text. The contributions of all co-authors in publications included in the body of the thesis have been declared, signed by each co-author and attached as an appendix. I hereby declare that I have not previously or concurrently submitted this material, either in full or in part, for a degree at this or any other institution.

Thaigarajan Parumasivam

Oct 2016

CONTENTS

Acknowledgement ...... viii

Glossary ...... x

Thesis abstract ...... 1

Chapter 1 ...... 1

Chapter 2 ...... 1

Chapter 3 ...... 2

Chapter 4 ...... 3

Chapter 5 ...... 4

Chapter 6 ...... 5

Chapter 7 ...... 6

Introduction ...... 7

Objectives and aims ...... 10

Chapter 1: Dry powder inhalable formulations for anti-tubercular therapy ...... 14

Abstract ...... 15

1. Introduction ...... 16

2. Powder formulation approaches using particle engineering techniques ...... 20

2.1 Formulations by spray drying ...... 35

2.1.1 Capreomycin sulphate ...... 36

2.1.2 Rifamycins (rifampicin and rifapentine) ...... 37

2.1.3 Para-aminosalicylic acid (PAS) ...... 39

2.1.4 Pretomanid (PA-824) ...... 40

iii

2.1.5 Isoniazid ...... 41

2.1.6 Combination formulations ...... 41

2.1.7 Vaccine candidates...... 42

2.1.8 Bacteriophage formulations ...... 43

2.2 Formulations by freeze drying and spray freeze drying ...... 46

2.3 Formulations by supercritical fluid precipitation ...... 47

3. Other inhalable dry powder production approaches ...... 48

3.1 Liposomal formulations ...... 48

3.2 Polymeric formulations ...... 50

4. Dry powder delivery devices ...... 53

5. Pre-clinical and clinical trial evaluations of inhaled TB drug powders ...... 57

5.1 In vitro studies...... 57

5.2 In vivo studies ...... 58

5.3 Clinical trial ...... 60

6. Risk related to inhaled regimen in the treatment of TB ...... 61

7. Expert opinion on future development ...... 64

8. Conclusions ...... 66

Acknowledgement ...... 67

References ...... 68

Chapter 2: In vitro evaluation of inhalable verapamil-rifapentine particles for tuberculosis therapy ...... 83

Abstract ...... 84

1. Introduction ...... 85

2. Materials and methods ...... 89

3. Results ...... 98

4. Discussion ...... 117

5. Conclusion ...... 122

Acknowledgement ...... 123

References ...... 124

Chapter 3: In vitro evaluation of novel inhalable dry powders consisting of thioridazine and rifapentine for rapid tuberculosis treatment ...... 128

Abstract ...... 129

1. Introduction ...... 130

2. Materials and methods ...... 134

3. Results ...... 142

4. Discussion ...... 156

5. Conclusion ...... 161

Acknowledgement ...... 162

References ...... 163

Chapter 4: Rifapentine-loaded PLGA microparticles for tuberculosis inhaled therapy: preparation and in vitro aerosol characterisation ...... 167

Abstract ...... 168

1. Introduction ...... 169

2. Materials and methods ...... 172

3. Results and discussion ...... 182

4. Conclusion ...... 203

Acknowledgement ...... 204

References ...... 205

Chapter 5: The delivery of high dose dry powder antibiotics by a low-cost generic inhaler

...... 209

Abstract ...... 210

1. Introduction ...... 211

2. Materials and methods ...... 216

3. Results ...... 225

4. Discussion ...... 239

5. Conclusion ...... 245

Acknowledgement ...... 246

References ...... 247

Chapter 6: Inhalation of respirable crystalline rifapentine particles induces pulmonary inflammation ...... 250

Abstract ...... 251

1. Introduction ...... 252

2. Materials and methods ...... 254

3. Results ...... 259

4. Discussion ...... 268

5. Conclusion ...... 273

Acknowledgement ...... 274

References ...... 275

Chapter 7: Conclusion, future studies and final remarks ...... 278

Conclusion ...... 279

Future studies ...... 281

Final remarks ...... 282

Appendix: Publications during candidacy, conference proceedings, supplementary data and co-author declaration ...... 283

Appendix 1: Publications during candidacy ...... 284

Appendix 2: Conference proceedings ...... 285

Appendix 3: Supplementary data from Chapter 4 ...... 294

Appendix 4: Supplementary data from Chapter 5 ...... 297

Appendix 5: Supplementary data from Chapter 6 ...... 302

Appendix 6: Co-author declaration ...... 306

vii

ACKNOWLEDGEMENT

“Gratitude unlocks the fullness of life. It turns what we have into enough, and more. It turns

denial into acceptance, chaos to order, confusion to clarity. It can turn a meal into a feast, a

house into a home, a stranger into a friend.”

- Melody Beattie

I wish to express my deepest gratitude to the following people for their admirable guidance, encouragement, and motivation for making this thesis a reality. It has been a privilege.

Foremost to my remarkable supervisory team – Hak-Kim Chan and Warwick Britton for imparting their knowledge and expertise, as well as their assistance in writing journals and reports. Both of them have given me endless guidance and hard questions not only in the scientific arena but also on a personal level. I will continue to look them as role models to success in my future career.

Very special thanks to John Chan and Sharon Leung for taking time out from their busy schedule to serve as external readers and well-wishers. I am also grateful to my lab mates;

Patricia Tang, Qi (Tony) Zhou, Anneliese Ashhurst, Angel Pang, Ariel Astudillo, Rachel

Chang, Sally Kim, Jennifer Wong, Siping (Alex) Sun, Dexter D’sa, Jiaqi Yu, and Michael

Yang for creating sweet memories to cherish and making life more amusing as well as meaningful in the lab. I offer my good luck and best wishes for their future undertakings. I

viii

would also like to thank Vicky Sifniotis and Ziya Sahin (office mates) for their exchange of knowledge, support and inspiration during tough times.

Not forgetting, none of this would have been possible without the love and patience from my family. Truthful gratitudes to my caring mom and dad, and understanding brothers (Kumar,

Manalan, Poovaneswaran, and Sathasivam) for their encouragement and tolerance with me all these years. You guys are always been there for me. No words can ever be expressed for my gratefulness and love for you all.

Last but not the least, this work would not have been possible without the funding provided by the Universiti Sains Malaysia and the Malaysian Government. I take this opportunity to express my sincere gratitude to them.

GLOSSARY

A549 Adenocarcinomic human alveolar basal epithelial cell

ADC Albumin, dextrose and catalase

ATCC American Type Culture Collection

BAL Bronchoalveolar lavage

BCG Bacilli Calmette-Guerin

CF Cystic fibrosis

CFD Computational fluid dynamics

CFU Colony forming unit

Cmax Peak concentration

COPD Chronic obstructive pulmonary disease

DMSO Dimetyl sulfoxide

DPI Dry powder inhaler

DPPC 1,2-dipalmitoyl-sn-glycero-3-phosphocholine

DSC Differential scanning calorimetry

DSPC Distearyl phosphatidylcholine

DVS Dynamic vapour sorption

FBS Fetal bovine serum

FDA Food and Drug Administration

FPF Fine particle fraction

FTIR Fourier transform infrared

GSD Geometric standard deviation

HIV Human immunodeficiency virus

HPLC High performance liquid chromatography

HPMC Hydroxypropyl methyl cellulose

HPSC Hydrogenated soya phosphatidylcholine

IC50 Half-maximal inhibitory concentration

IL Interleukin

Leu L-leucine

LOQ Limit of quantification

MDI Metered dose inhaler

MDR-TB Multi-drug resistant tuberculosis

MIC90 Minimum drug concentration at which 90 % of bacteria inhibited

MMAD Mass median aerodynamic diameter

MOI Multiplicity of infection

MSLI Multi-stage liquid impinger

MW Molecular weight

NGI Next generation impactor

O/W Oil in water phase

OADC Oleic acid, albumin, dextrose and catalase

PBS Phosphate buffer saline

PLA Poly (lactic acid)

PLGA Poly (lactic-co-glycolic) acid pMDI Pressurised metered dose inhaler

PVA Polyvinyl alcohol

RH Relative humidity

RI Refractive index

RPT Rifapentine

SEM Scanning electron microscopy

TB Tuberculosis

TDR-TB Totally drug-resistant tuberculosis

Tg Glass transition temperature

TGA Thermogravimetric analysis

TGA Therapeutic Good Administration

THP-1 Human monocytic cell line

THZ Thioridazine

Tmax Peak time

TNF Tumor necrosis factor

TLR Toll like receptor

USP United States Pharmacopeia

VMD Volumetric median diameter

VRP Verapamil

W/O/W Water in oil in water double emulsion

WHO World Health Organization

XDR-TB Extensively drug-resistant tuberculosis

XRPD X-ray powder diffractometry

xii

THESIS ABSTRACTS

Chapter 1

Inhalable drugs formulated as dry powders are more stable than the liquid dosage forms, do not require refrigeration, and are more efficient for the delivery of large doses likely required for anti-tubercular drugs. This chapter contains a literature review examining published techniques for producing inhalable anti-tuberculosis (anti-TB) dry powder formulations. The review also covers the current innovations of dry powder inhalers (DPIs) and potential risk of delivering therapeutic drugs directly to the lungs.

This chapter has been published in Advanced Drug Delivery Reviews, 2016; 102: 83-101, with the title ‘Dry powder inhalable formulations for anti-tubercular therapy’. Authors:

Parumasivam T, Chang RYK, Abdelghany S, Ye TT, Britton WJ, Chan H-K.

Chapter 2

Efflux pumps are broadly recognised as one of the key contributors to the development of drug resistance in Mycobacterium tuberculosis. It has been demonstrated in preclinical studies that co-administration of drugs that inhibit these efflux pumps (efflux pump inhibitors) with the anti-TB regimen could significantly enhance the efficacy of the treatment. Verapamil is an example of an efflux pump inhibitor which appears to be promising adjunctive anti-TB therapy.

The drug has also been approved by the U.S. Food and Drug Administration (FDA) for the treatment of hypertension. Verapamil, however, has poor bioavailability when administered

orally with rifamycins (i.e. rifampicin and rifapentine), thereby limiting its clinical effectiveness. This is because rifamycins induce the hepatic enzyme CYP3A4 which metabolises verapamil in the liver. To overcome this limitation, an inhalable dry powder consisting of amorphous verapamil and crystalline rifapentine was produced. The powder was found suitable for aerosol delivery with a respirable fraction (< 5 µm) greater than 70 %. The combination also showed higher killing of M. tuberculosis within the macrophages compared to either verapamil or rifapentine alone. The combination showed an acceptable half maximal inhibitory concentration (IC50) of 62.5 µg/mL on both human monocytic (THP-1) and lung alveolar basal epithelial (A549) cell lines. The powder required a moisture-free packaging for long-term storage stability.

This chapter has been published in Molecular Pharmaceutics, 2016; 13: 979-989, with the title

‘In vitro evaluation of inhalable verapamil-rifapentine particles for tuberculosis therapy’.

Authors: Parumasivam T, Chan JGY, Pang A, Quan DH, Triccas JA, Britton WJ, Chan H-K.

Chapter 3

The next publication explores the development of an inhalable dry powder consisting of thioridazine and rifapentine. Similar to verapamil, thioridazine is an efflux pump inhibitor that has potential in the treatment of tuberculosis (TB). Thioridazine is an FDA-approved antipsychotic drug. However, the drug can induce adverse cardiac effects in TB patients when used at an efficacious oral dose of 200 mg/day. The total dose could be reduced via an inhaled route of delivery while increasing pulmonary drug concentrations. Inhalable dry powders consisting of thioridazine and rifapentine were produced: (I) Formulation 1 is a combination of amorphous thioridazine and crystalline rifapentine, and (II) Formulation 2 containing both drugs in amorphous forms. Both powder formulations were found suitable for aerosol delivery with a 2

respirable fraction (< 5 µm) between 68 and 76 %. Formulation 1, however, exhibited greater chemical stability after storage for 3 months at a low humidity (vacuum) condition. The thioridazine and rifapentine in combination showed better bactericidal activity on M. tuberculosis within the macrophages compared to the single drug treatments. In regards to cytotoxicity, the combination exhibited an IC50 of 31.25 µg/mL on both THP-1 and A549 cell lines.

This chapter has been published in European Journal of Pharmaceutics and Biopharmaceutics,

2016; 107: 205-214, with the title ‘In vitro evaluation of novel inhalable dry powders consisting of thioridazine and rifapentine for rapid tuberculosis treatment’. Authors: Parumasivam T,

Chan JGY, Pang A, Quan DH, Triccas JA, Britton WJ, Chan H-K.

Chapter 4

In the following chapter, the potential of polymeric particles for aerosol delivery of anti- tubercular drugs was investigated. Poly(lactic-co-glycolic acid) (PLGA) is an FDA-approved polymer for medical application because of its biocompatibility and biodegradation. However, limited data are available on the effects of monomer ratio (lactic and glycolic acid) and molecular weight (MW) on the aerosol performance of the powder. The suitability of PLGAs for delivering rifapentine was assessed by evaluating PLGAs with monomer ratios of 50:50,

75:25 and 85:15, and molecular weight (MW) ranging between 24 and 240 kDa. Regardless of the monomer ratio and MW, the particles were within the inhalable size of less than 5 µm, and the fine particle fraction (FPF) (particles with an aerodynamic diameter < 5 µm relative to the total recovered drug mass) ranged between 52 and 57 %. These powders were also non-toxic to THP-1 and A549 cells at concentrations up to 1.5 mg/ml. PLGA 85:15 particles exhibited the highest likelihood of phagocytosis by THP-1 monocyte-derived macrophages, followed by 3

75:25 and 50:50. These findings suggest that the monomer ratio and MW of PLGA did not have any significant effect on the aerosol performance and toxicity, yet may have an indirect influence on enhancing the killing of the intracellular M. tuberculosis.

This chapter has been published in European Journal of Pharmaceutical Sciences, 2016; 88: 1-

11, with the title ‘Rifapentine-loaded PLGA microparticles for tuberculosis inhaled therapy: preparation and in vitro aerosol characterization’. Authors: Parumasivam T, Leung SSY, Quan

DH, Triccas JA, Britton WJ, Chan H-K.

Chapter 5

Many capsule-based DPIs have been optimised to deliver a small amount of drugs, generally less than 20 mg, for the treatment of asthma and chronic obstructive pulmonary disease

(COPD). When delivering higher doses using these inhalers, patients need to load and inhale multiple capsules, resulting in a tedious dose administration process. To overcome the requirement of loading multiple capsules, an Aerolizer® dry powder device was modified to accommodate a size 0 capsule for delivery of high doses to the lungs in a single capsule with multiple inhalation manoeuvres. In addition to a large capsule size, in some devices, the four piercing pins (0.6 mm diameter) were replaced with a single centrally located 1.2 mm pin and

1/3 reduced air inlet relative to full inlet of the original design. The in vitro aerosol performance of rifapentine and another two drugs, viz. colistin and tobramycin, was evaluated at capsule loadings of 30, 50 and 100 mg, respectively. The device with a single pin and 1/3 air inlet showed superior performance dispersing rifapentine and colistin. Subsequently, a device with a similar configuration but with an expanded capacity to accommodate a size 00 capsule was manufactured to increase the drug payload further. When the device was tested at inhalation conditions achievable by cystic fibrosis (CF) patients, a capsule loading of 100 mg rifapentine 4

and colistin showed optimal performance at 60 L/min with an air volume of 2.0 L and 45 L/min with an air volume of 1.5 – 2.0 L, respectively. These results demonstrate the potential of the modified Aerolizer® device to deliver high doses of dry powder systems including anti- tubercular drugs.

This chapter has been published in The AAPS Journal, 2016; 19: 191-202, with the title ‘The delivery of high dose dry powder antibiotics by a low-cost generic inhaler’. Authors:

Parumasivam T, Leung SSY, Tang P, Mauro C, Britton WJ, Chan H-K.

Chapter 6

The penultimate chapter of the thesis addresses the potential in vivo inflammatory risk of delivering rifapentine particles directly to the lungs. A single dose of 20 mg/kg rifapentine dry powder was delivered to 6 - 8-week old BALB/c mice by intratracheal insufflation. The mice were euthanised at 12 h, 24 h and 7 days post-treatment, and the lungs and bronchoalveolar lavage (BAL) fluid were collected. The control mice were untreated (Nil). The lung and BAL were processed for total cell counts and immune cell characterisation via flow cytometry. One right lobe of the lung was fixed in 10 % neutral buffered formalin, and stained with hematoxylin and eosin (H&E) for histopathological examination. Pulmonary delivery of rifapentine caused a significant influx of neutrophils in the lungs by 12 h that had resolved over 7 days. In the

BAL, there was a significant increase in the abundance of alveolar macrophages at 24 h, and this also resolved by day 7. The histopathological patterns were consistent with these findings: pulmonary congestion and a massive increase in the abundance of macrophages was observed at 12 and 24 h post-treatment. In conclusion, pulmonary delivery of rifapentine particles caused a transient neutrophil-associated inflammation in the lungs that resolved over 7 days. This

finding suggests that inhalation of rifapentine is suitable for delivering once a week at a dose of 20 mg/kg or less.

This chapter has been published in Molecular Pharmaceutics, 2017; 14: 328-335, with the title

‘Inhalation of respirable crystalline rifapentine particles induces pulmonary inflammation’.

Authors: Parumasivam T, Ashhurst AS, Nagalingam G, Britton WJ, Chan H-K.

Chapter 7

The major findings from the previous chapters are summarised in this chapter.

 Novel inhalable dry powders consisting of efflux pump inhibitors (verapamil

and thioridazine) and rifapentine were successfully formulated and were shown

to be suitable for aerosol administration.

 Inhalable polymeric PLGA particles containing rifapentine were shown suitable

for targeting intracellular M. tuberculosis and were relatively non-toxic to

respiratory cells.

 A modified Aerolizer® dry powder inhaler device which can accommodate a

size 00 capsule was shown capable of aerosolising high dose dry powder

antibotics at doses up to 100 mg in a single capsule.

 Pulmonary delivery of rifapentine caused a neutrophil-associated inflammatory

response in the lungs that had resolved over 7 days.

Finally, future projects and challenges that accelerate the progress of inhaled anti-TB therapy are discussed.

INTRODUCTION

It is estimated that one-third of the global population has been infected with Mycobacterium tuberculosis, a deadly lung pathogen. Approximately, 10 % of infected individuals will develop active tuberculosis (TB). Unlike common respiratory infections that typically require one to two weeks of oral antibiotics, successful treatment of TB requires at least six months of multiple oral antibiotics administered daily. This lengthy treatment duration leads to poor patient adherence and severe side effects which has fuelled the development of multi-drug resistant (MDR) and extremely drug resistant (XDR) strains of M. tuberculosis. Therefore, there is an urgent need for an improved therapeutic option that can address these limitations in treating TB infection.

Rifapentine is an antibiotic approved by the U.S. Food and Drug Administration (FDA) since

June 1998 for the treatment of pulmonary TB and is administered orally. Recent murine preclinical efficacy studies suggest that if the currently approved dose of 10 mg/kg twice- weekly was increased to 20 mg/kg daily, TB can be treated in just three months or less [1].

However, these shorter treatment times were not replicable in the human clinical trial [2]. This was attributed to a multitude of physiological barriers when rifapentine is administered orally in human, including poor absorption and liver metabolism that led to low systemic drug concentrations and limited access of rifapentine to the lungs [2, 3]. It is hypothesised that if rifapentine is delivered directly to the pulmonary site of infection, these limitations could be overcome and the treatment duration can be shortened.

The pulmonary route of drug delivery, which mimics the natural route of M. tuberculosis infection, has attracted significant attention for treatment of TB as it offers several advantages compared to systemic delivery. These including: (i) delivering high drug concentrations to the lungs, (ii) avoiding hepatic first-pass metabolism, (iii) potentially the slow release in the lungs may reduce the dosing frequency, and (iv) directly targeting the primary site of infection that harbours M. tuberculosis. Indeed, inhaled antibiotics have established clinical efficacy in treating a variety of respiratory infections, such as CF and pneumonia. However, no inhaled medications are currently approved for treatment of TB. Nonetheless, anti-TB drugs administered via the pulmonary route have demonstrated favourable pharmacokinetic and pharmacodynamic in a number of preclinical studies, and in one Phase 1 study with inhaled capreomycin [4].

The nebuliser, pressurised metered-dose inhaler (pMDI) and dry powder inhaler (DPI) are popular models of pulmonary delivery devices. The nebuliser generates aerosol medication from liquid or suspension formulations in the form of mist using compressed gas or ultrasonic waves. The aerosolisation mechanism of pMDI is driven by propellants, such as chlorofluorocarbon or hydrofluorocarbon. The drug is dissolved in the propellant under pressure and releases a metered volume of drug-propellant suspension when activated. DPI is a breath-activated device ideal for dry powder delivery. The powder is dispersed into inhalable particles by the inspiratory flow generated by the user. For treatment of TB, patients may need to inhale a large total dose (over a hundred milligrams) of an antibiotic for an effective clinical outcome. pMDI delivers microgram doses of drugs and requires hand-breath coordination, making them unsuitable. While the nebuliser can deliver a high dose, the process is time- consuming and the devices are relatively bulky and require a source of electrical power.

Currently, DPIs are designed for delivering tens of milligrams of powders and so with the

higher doses likely to be required for TB treatment, patients may need to go through more tedious multi-dose inhalations. Therefore, device enhancements and/or design of tailored devices for high powder doses need to be developed to ensure that inhaled dry powder antibiotic therapy for TB is viable. Furthermore, dry powders have superior chemical stability and bacterial resistance compared to liquid dosage forms.

Although inhaled therapy holds potential for rapid treatment of TB, there are also some risks that need to be balanced. For example, direct deposition of particles in the already inflamed lungs of the patients could exacerbate the pulmonary inflammation and lung damage.

Therefore, these risks should be evaluated during the inhaled product development.

OBJECTIVES AND AIMS

The core objective of the studies covered in this thesis is to demonstrate the potential and any related challenges of inhaled therapy for the treatment of TB using the inhalation of rifapentine as a model. The thesis embarks on the following specific objectives:

1. To summarise the recent advances in inhaled dry powder formulations, including

vaccine and drug candidates for treatment and prevention of TB.

2. To investigate the limitations of the current oral rifapentine TB treatment, and provide

reasoning to develop inhalable drug powders.

3. To produce novel inhalable dry powders containing rifapentine either as pure drug or

in combination with other drugs or with carriers, such as polymers, to overcome the

limitations identified in Objective 2.

4. To develop a novel DPI that can deliver high doses of rifapentine and other drugs.

5. To assess the murine inflammatory response when dry powder rifapentine is delivered

to the lungs

The overall research flow to achieve these objectives is outlined in Figure 1. Chapter 1 summarises the key findings and updates on the dry powder formulations for treatment and prevention of TB. It also covers the current development in the DPI designs and challenges pertaining to inhaled delivery in the context of TB treatment. Chapters 2 and 3 are about two novel formulations consisting of rifapentine in combination with efflux pump inhibitor drugs which hold a high potential to prevent the emergence of drug resistance during treatment and

hence shorten the duration of anti-TB regimen. Chapter 4 examines the potential of inhaled rifapentine-loaded PLGA particles for targeting intracellular M. tuberculosis. The following

Chapter 5, evaluates a modified Aerolizer® device as a low-cost generic inhaler for delivery of high dose dry powder antibiotics, including rifapentine. The final chapter (Chapter 6) examines the pulmonary inflammatory response to respirable rifapentine particles in murine toxicity studies.

Inhaled therapy for rapid tuberculosis treatment

Novel Dry powder formulations

Chapter 1 - Review

Antibiotic dry powder Polymeric dry powder

Chapter 2 – Verapamil-rifapentine Chapter 4 – PLGA-rifapentine Chapter 3 – Thioridazine-rifapentine

High dose delivery device Chapter 5: Modified Aerolizer®

Pulmonary toxicity

Chapter 6: Pulmonary toxicity of inhaled rifapentine

Conclusion and future studies

Figure 1. Flowchart of research progress.

REFERENCES

[1] I.M. Rosenthal, R. Tasneen, C.A. Peloquin, M. Zhang, D. Almeida, K.E. Mdluli, P.C. Karakousis, J.H. Grosset, E.L. Nuermberger, Dose-ranging comparison of rifampin and rifapentine in two pathologically distinct murine models of tuberculosis, Antimicrob Agents Chemother, 56 (2012) 4331-4340. [2] K.E. Dooley, E.E. Bliven-Sizemore, M. Weiner, Y. Lu, E.L. Nuermberger, W.C. Hubbard, E.J. Fuchs, M.T. Melia, W.J. Burman, S.E. Dorman, Safety and pharmacokinetics of escalating daily doses of the antituberculosis drug rifapentine in healthy volunteers, Clin Pharmacol Ther, 17 (2012) 521-525. [3] K.E. Dooley, R.M. Savic, J.-G. Park, Y. Cramer, R. Hafner, E. Hogg, J. Janik, M.A. Marzinke, K. Patterson, C.A. Benson, L. Hovind, S.E. Dorman, D.W. Haas, A.A.S. Team, Novel dosing strategies increase exposures of the potent antituberculosis drug rifapentine but are poorly tolerated in healthy volunteers, Antimicrob Agents Chemother, 59 (2015) 3399- 3405. [4] A.S. Dharmadhikari, M. Kabadi, B. Gerety, A.J. Hickey, P.B. Fourie, E. Nardell, Phase I, single-dose, dose-escalating study of inhaled dry powder capreomycin: a new approach to therapy of drug-resistant tuberculosis, Antimicrob Agents Chemother, 57 (2013) 2613-2619.

CHAPTER 1

DRY POWDER INHALABLE FORMULATIONS FOR ANTI-

TUBERCULAR THERAPY

This chapter has been published in Advanced Drug Delivery Reviews, 2016; 102: 83-101, with the title ‘Dry powder inhalable formulations for anti-tubercular therapy’ Authors:

Parumasivam T, Chang RYK, Abdelghany S, Ye TT, Britton WJ, Chan H-K.

ABSTRACT

Tuberculosis (TB) is an infectious disease caused by the airborne bacterium, Mycobacterium tuberculosis. Despite considerable research efforts, the treatment of TB continues to be a great challenge in part due to the requirement of prolonged therapy with multiple high-dose drugs and associated side effects. The delivery of pharmacological agents directly to the respiratory system, following the natural route of infection, represents a logical therapeutic approach for treatment or vaccination against TB. Pulmonary delivery is non-invasive, avoids first-pass metabolism in the liver and enables targeting of therapeutic agents to the infection site. Inhaled delivery also potentially reduces the dose requirement and the accompanying side effects. Dry powder is a stable formulation of drug that can be stored without refrigeration compared to liquids and suspensions. The dry powder inhalers are easy to use and suitable for high-dose formulations. This review focuses on the current innovations of inhalable dry powder formulations of drug and vaccine delivery for TB, including the powder production method, preclinical and clinical evaluations of inhaled dry powder over the last decade. Finally, the risks associated with pulmonary therapy are addressed. A novel dry powder formulation with high percentages of respirable particles coupled with a cost effective inhaler device is an appealing platform for TB drug delivery.

1. INTRODUCTION

Tuberculosis (TB) is a chronic infectious disease caused by Mycobacterium tuberculosis that results in more deaths than any other communicable disease [1, 2]. Recently, the World Health

Organization (WHO) declared 1.5 million deaths globally out of an estimated 9 million people who developed TB in 2013 [1]. The threat is exaggerated by co-infection with the human immuno-deficiency virus (HIV) and the emergence of multi-drug-resistant (MDR) and extensively drug-resistant (XDR) strains of M. tuberculosis. In 2013, 1.1 million TB cases were reported among HIV-positive patients, 480 000 new cases of MDR-TB and nearly 50 000 people with XDR-TB [1]. These numbers may be underestimated due to lack of drug-resistance testing and treatment facilities across high-risk regions, such as sub-Saharan Africa, Asia, and

Eastern Europe [3].

M. tuberculosis is transmitted predominantly through small airborne droplets or droplet nuclei created by the coughing or sneezing of a person with pulmonary or laryngeal TB. Once the M. tuberculosis in droplet nuclei reaches the terminal airways, they are quickly surrounded and engulfed by alveolar macrophages. The bacterial cells within the macrophages are then exposed to multiple microbicidal mechanism including reactive oxygen and nitrogen species, acidic pH, lysosomal enzymes, and toxic peptides [4, 5]. Furthermore, the macrophages initiate production of proteolytic enzymes and cytokines to degrade these pathogenic microorganisms.

The release of cytokines consequently attracts T lymphocytes to the infected site [6]. In an immunocompetent person, granulomas will be formed through accumulation of activated T lymphocytes and macrophages around the M. tuberculosis to contain the infection [4, 7]. These tumour necrosis factor (TNF)-dependent granulomas create a microenvironment that limits the replication and spreading of M. tuberculosis [8, 9]. This condition results in latent TB infection

(LTBI), where the infection is not active and cannot be transmitted [10, 11]. Nonetheless, these dormant cells can be reactivated when the host immune system is suppressed (e.g. by HIV infection) [11, 12]. On the other hand, in immunocompromised patients the granuloma formation fails to trap or contain the bacteria [13]. Therefore, the M. tuberculosis will be able to escape and spread to other alveoli and organs that lead to progressive primary TB [14].

Until the present, the only approved TB vaccine is an attenuated form of Mycobacterium bovis, bacilli Calmette-Guerin (BCG) developed by Calmette and Guérin in 1908 [15]. BCG is a live attenuated strain of M. bovis that was isolated from a cow with tuberculous mastitis [15]. BCG vaccination has been successfully carried out in children five years and under for over a century. The vaccine is relatively cheap and safe, and is administered as a single intradermal injection. BCG is protective against childhood and disseminated TB, but less effective against adult pulmonary TB [16, 17]. In addition, in clinical trials the efficacy of BCG demonstrated the protection wanes over time and only lasts up to 10 years after vaccination [18, 19]. A number of meta-analyses have further revealed that the efficacy of BCG immunisation for prevention of adult pulmonary TB depends on the geographical location where the vaccination was carried out [17, 20, 21]. BCG confered a strong protection against TB to people living in the UK (77 %), but was less efficacious in South India, suggesting the host response to BCG is strongly modulated by environmental and genetic factors [16]. As a result, the protective efficacy of BCG has varied dramatically between different parts of the world and its impact on the global control of TB remains ambiguous [17, 20, 21].

The WHO recommended TB therapy requires large doses of antibiotics in combination to be taken orally every day for at least six months. In general, during the initial phase the patients receive three to four first-line drugs (isoniazid, rifampicin, pyrazinamide, and ethambutol)

daily for two months for a rapid amelioration of clinical symptoms [22, 23]. In the continuation period, two to three drugs (isoniazid, rifampicin or ethambutol) are given intermittently for four to seven months to eliminate the remaining M. tuberculosis cells and reduce the relapse rate [22, 23]. The current treatment of MDR-TB requires at least eight months for the initial phase with a regimen containing 2 - 3 injectable second-line drugs, such as aminoglycosides

(kanamycin or amikacin) and polypeptides (capreomycin or viomycin), and then three oral drugs including a fluroquinolone, ethambutol and possibly ethionamide for up to 20 months

[24]. The treatment of XDR-TB is substantially longer (> 12 months) and requires administration of the third-line anti-TB drugs (clofazimine, linezolid, amoxicillin plus clavulanate and imipenem plus cilastatin), which are expensive and more toxic [24, 25]. The long duration of medication and unpleasant side effects of the drugs often lead to high rates of noncompliance with the consequence of increased mortality rate. The recent emergence of totally drug-resistant TB (TDR-TB) cases has exacerbated the problem. Hence, there is a strong demand for a simpler, affordable and more effective medication approach for treatment of TB.

Over the past decade, there have been increasing research interests towards pulmonary route for treatment and prevention of TB [21, 26-28]. Compared to conventional administrations, the advent of inhalation as a potential route for treatment of TB has the following advantages for the second and third line drugs:

(i) Noninvasive therapy avoids the necessity of injections for second and third-line

anti-TB drugs and avoids the requirement for the disinfection of injection sites

and disposal of the needles [29],

(ii) Allows a high drug concentration in the lung which can reduce the frequency of

dosing [29, 30],

(iii) May require smaller inhaled doses compared to those required for oral route

[28],

(iv) Targets alveolar macrophages which harbour M. tuberculosis cells [31-34],

(v) Activates the innate bactericidal machinery of infected macrophages [35],

(vi) Aerosol vaccination may best mimic the immune response in the pulmonary

region followed by systemic immunity [27, 36-39], and

(vii) Direct absorption of active agents into systemic circulation bypassing hepatic

first-pass metabolism could avoid the degradation of drug or vaccine by the

acidic environment in the stomach [40].

Despite these advantages compared to the current anti-TB regimen, no commercial inhaled medication is currently approved for treatment of TB, and this may be due to the following factors:

1. Issues in dose estimation and dosing reproducibility as many parameters could

affect the lung deposition, i.e. breathing pattern, relative humidity (RH), particle

morphology, surface, density or shape, and inhaler device design [41, 42],

2. Inhaled medication and inhaler devices are potentially more costly, and may not

be affordabable in low-income countries [41],

3. Lack of trained staff to instruct on the proper use of devices for administration

[42], and

4. Insufficient patient compliance and adherance to new inhaled treatments [43].

5. While most of these issues can be addressed ultimately, in the authors’ own

opinions, there are other hurdles which may be more important: 1. TB

physicians may not utilise inhaled therapy owing to the uncertainty of whether

(i) the inhaled drugs will penetrate into the site of infection, including into

granulomas, and (ii) delivery of inhaled drugs into already inflammed lungs

would cause unwanted adverse effects. 2. Global organisations for TB treatment

may favour oral administration and new drug discovery rather than repurposing

old drugs for an alternative inhalation route.

This review mainly focuses on recent advances in inhaled dry powder formulations, including drug and vaccine candidates for treatment and prevention of TB. Issues and challenges pertaining to the development of dry powder formulation with promising aerosol performance and long-term physicochemical stability are addressed. In addition, the current development in the inhaler device design for delivering a high dose antibiotic amongst TB patients is covered.

Finally, the clinical and safety aspects of dry powder inhalation are discussed.

2. POWDER FORMULATION APPROACHES USING PARTICLE ENGINEERING

TECHNIQUES

Dry powder formulations may offer greater physiochemical stability over liquid or suspension- based formulations. Moreover, they do not require refrigeration and are easy to transport. In general, the powder requires moisture-free packaging to protect them from humidity (and light) for long-term storage. For inhalation, particles aerodynamically within 1 - 5 µm in size are considered suitable for deposition in the lungs.

The traditional way of producing dry powders for aerosol delivery involves crystallisation, micronisation through milling, followed by blending with coarse carrier particles [44-46].

Lactose is a commonly used carrier for improving the flowability of a powder mixture for better capsule emptying and enhanced deposition of therapeutic particles in the lungs [47]. A typical

drug-to-carrier proportion is 1:67.5 [48]. Because of the large dose likely to be required for inhaled treatment for TB, the lactose carrier system may not be suitable. As a result, alternative methods, such as spray drying and spray freeze drying have been utilised. These techniques are advantageous in terms of easy production scalability and better control over particle size distribution for enhanced aerodynamic performance [44]. Other approaches involving liposomal and polymeric formulations may improve the physicochemical stability and offer protection for the active ingredient, although the cost can be a barrier for large-scale production.

Tables 1 and 2 summarise the key findings of dry powder formulations of anti-TB drugs and vaccines, respectively.

Table 1: Summary of inhaled TB antibiotic powder formulations with the processing parameters and the key findings.

Drug Production Excipient Formulation description Inhaler Flow FPF (%) Key findings Refs method device rate (L/min) Amikacin Freeze drying Hydrogenated Liposomal formulation Rotahaler 30, 60 21.9-34.2 Addition of lactose carrier was [49] sulfate soybean and 90 crucial to enhance the phosphatidylcholine, flowability of liposome powder cholesterol saturated soybean phosphatidylglycerol or stearylamine and lactose carrier

Spray drying Hydrogenated soya Liposomal formulation Rotahaler 28.3 Spray dried, Spray dried particles found to [50] (inlet, 125- phosphatidylcholine, 65.1; have better aerosol 130°C; outlet, leucine and freeze performance compared to 65-70°C; fd, Poloxamer 188 dried, 50.5 freeze dried powder due to less 2mL/min; asp, (< 5 µm) device and capsule retention 100 %, atom, 1.5 kg/cm2) and freeze drying

Capreomycin Spray drying Leucine Large porous particles A hand- 28.3 50 (< 5.8 Longer half-life in guinea pigs [29, sulfate (inlet, 186- held, dry µm) and 10 51, 189°C; outlet, powder, (< 3.3 µm) Significantly reduced the 52] 65°C; fd, 80 breath- bacterial count in the lung of mL/min; asp, activated guinea pigs challenged with M. 79-82 kg/h; inhaler tuberculosis atom, 29-31 device g/min) Inhalation of single dose 300 mg achieved serum concentration above MIC of M. tuberculosis and the dose well tolerated in humans

Supercritical Leucine Crumpled particles PuffHaler 28.3 > 21.6 (< Contained low moisture [53] precipitation and and 60 3.3 µm) content and retained the anti- and spray drying Aerolizer TB activity after processed (inlet, 60°C; asp, 30 L/min)

Nano spray Lactose Spherical particles with a HandiHaler 60 11-30 Drug particles needed lactose [54] drying (inlet, smooth surface and to improve respirability 100-120°C; smaller size atom, 100 %)

Frozen in liquid PLGA, leucine and Double emulsion (w/o/w) Spinhaler 100 54.3 (< 5 Encapsulated drug showed [55] nitrogen and mannitol and polymeric particle µm) enhanced killing compared to freeze drying free drug treatment in M. tuberculosis infected THP-1 cells

Freeze drying PLGA Double emulsion (o/w/o) n/a n/a n/a The method was not easy to [56] and porous polymeric scalable and the powder microparticles required further analysis on aerosol performance

Spray drying Oleate, linoleate and Hydrophobic ion-pair and n/a n/a n/a Capreomycin oleate was less [57] (inlet, 70-80°C; linolenate nanopowder toxic and showed comparable fd, 2 mL/min; antimycobacterial activity to asp, 27 m3/h; capreomycin alone atom, 414 L/h) and nano spray drying (inlet, 80-90°C; fd, 1; spray, 100 %; asp, 100 L/min)

Clofazimine Spray drying Leucine None n/a n/a n/a Treatment of infected mice [58] (n/a) resulted in significant reduction in bacterial count than the oral administration

Isoniazid Supercritical Leucine Spheres, crystals and PuffHaler 28.3 > 21.6 (< Contained low moisture [53] precipitation some cube-like particles and and 60 3.3 µm) content and intact anti-TB and spray drying Aerolizer activity after processed (inlet, 60°C; asp, 30 L/min)

Spray drying L-α soybean Proliposome formulation A dispersion 60 15-35 (< Powder was non-toxic to [59] (inlet, 110°C; phosphatidylcholine, with mannitol as carrier device made 4.4 µm) respiratory cells including outlet, 88-92°C; cholesterol and in-house human epithelial cells and fd, 3 mL/min; mannitol alveolar macrophages and did atom, 800 kPa) not activate inflammatory mediators of alveolar macrophages

Spray drying Chitosan, Chitosan/tripolyphosphate Cyclohaler 28.3 67 High FPF with sustained drug [60] (inlet, 140°C; tripolyphosphate, particles release profile outlet, 70-80°C; lactose and leucine fd, 5 mL/min; asp, 45 m3/h; atom, 3 kg/cm2)

Spray drying Chitosan and Chitosan/tripolyphosphate Cyclohaler 60 7-45 Nanoparticles showed lower [61] (inlet, 110°C; tripolyphosphate, nanoparticles MIC than the free isoniazid fd, 20 %; asp, with lactose, against Pseudomonas 80 %; atom, 600 mannitol and aeruginosa and Staphylococcus L/h) maltodextrin alone aureus. Nanoparticles blended or with leucine with lactose and leucine showed highest FPF

Kanamycin Supercritical Leucine Crumpled particles PuffHaler 28.3 > 21.6 (< Contained low moisture [53] precipitation and and 60 3.3 µm) content and retained the drug and spray drying Aerolizer anti-TB activity (inlet, 60°C; asp, 30 L/min)

Levofloxacin Spray drying L-α soybean Proliposome formulation A dispersion 60 13-38 (< Proliposomes had enhanced [62] (inlet, 90°C; fd, phosphatidylcholine, with porous mannitol as device made 4.4 µm) activity against M. bovis than 3 mL/min; cholesterol and carrier in-house the free levofloxacin. No renal atom, 800 kPa) porous mannitol and liver toxicity was observed in the rats repeatedly treated with levofloxacin-liposomes

PA-824 Spray drying Leucine and Porous particles Hand-held, n/a 53.3 (< 5.8 Drug concentration in the lungs [63, (inlet, 107°C; phospholipid 1,2- breath- µm) of guinea pigs treated with 64] fd, 60 mL/min; dipalmitoyl-sn- activated, pulmonary route showed drug atom, 25 g/min) glycero-3- capsule- concentration that remained for phosphocholine based dry 32 h post-exposure, whereas powder those given the drug orally inhaler cleared the drug more rapidly device Infected animals treated with inhalation of PA-824 showed a lower number of viable bacteria than the untreated group. Greater bacterial reduction was reported in the lungs and spleen of animals given oral treatment

Para- Spray drying Ammonium Irregular crystal Spinhaler 60 24.5 (< The lower inlet temperature [65] aminosalicylic (inlet, 78°C; carbonate agglomerates to spherical 6.4 µm) (78°C) prevented the acid (PAS) outlet, 42-55°C; crystal clusters and conversion of PAS into acidic fd, 80 mL/min; porous microparticles form during spray drying asp, 100 %; atom, 570 NI/h) 25

Spray drying 1,2-Dipalmitoyl-sn- Large porous particles AIR 60 59-63 (< Rapid systemic drug onset and [66] (n/a) glycero-3- 5.6 µm), high concentration in the lung phosphocholine 31-39 (< lining fluid rats 3.4 µm)

Frozen in liquid PLGA, leucine and Single emulsion (o/w) Spinhaler 100 54.3 (< 5 No significant difference [55] nitrogen and mannitol and polymeric particle µm) between encapsulated and free freeze drying drug treatment in M. tuberculosis infected THP-1 cells

Pyrazinamide Spray drying Soybean Proliposome formulation A dispersion 60 20-30 (< Proliposomes were less toxic to [67] (inlet, 90°C; fd, phosphatidylcholine, with porous mannitol as device made 4.4 µm) respiratory cell lines and did 3 mL/min; cholesterol, and carrier in-house not activate inflammatory atom, 800 kPa) porous mannitol response in alveolar macrophage

Rifampicin Supercritical Lactose Spherical and n/a n/a n/a No chemical reaction between [68] precipitation polymorphic mixture rifampicin and lactose (CO2 flow, 20 g/min; CO2 temp, 38-42°C) and spray drying (inlet, 38-42; fd, 1 mL/min; asp, 2 g/min)

Recrystallisation None A very thin flaky Aerolizer or 60 HandiHaler, Flaky morphology reduced the [69] and spray drying structure and crystalline HandiHaler 63.8; agglomeration tendency and (inlet, 70°C; fd, Aerolizer, chemically stable up to 9 5 mL/min; asp, 68.5 (< 5 months 40 m3/h; atom, µm) 500 L/h)

Recrystallisation PLA or PLGA Thin flaky Rifampicin Aerolizer 60 24-45 (< 5 The coated rifampicin [70] and spray drying dihydrate coated with µm) dihydrates showed slower (inlet, 70°C; PLA or PLGA release than the uncoated. No outlet, 40-50°C; significant chemical fd, 8 mL/min; degradation in the rifampicin asp, 40 m3/h; crystals compared to atom, 600 L/h) amorphous form

Spray drying Chitosan Porous and wrinkled n/a 60 58-62 (5 Sustained drug release (Tmax 8 [71] (inlet, 150°C; particles µm) h) in vivo animal study fd, 5 mL/min; atom, 2.5 kg/cm2)

Spray drying Chitosan and Wrinkle surfaced porous Aerolizer 60 51.5-72.7 (5 Inhalac 230 improved the [72] (inlet, 150°C; hydroxyl propyl microparticles with µm) dispersibility of powder on fd, 5 mL/min; methyl cellulose Inhalac 230 as a carrier inspiration atom, 2.5 (HPMC) and kg/cm2) blended with Inhalac 230

Freeze drying PLGA Single emulsion (w/o)- A Model 30 46.7-69.9 Premix membrane [73, solvent evaporation with DP-4 dry (1-5 µm) homogenisation can be used to 74] premix membrane powder prepare monodisperse particles homogenisation PLGA insufflatorTM with extended drug release and microspheres (for rat) suitable aerodynamic performance for intratracheal delivery in rats

Freeze drying L-α soybean and Proliposome formulation A dispersion 60 Mannitol, Mannitol as a cryoprotectant [75, phosphatidylcholine, device made 66.8; produced crystalline powder 76] cholesterol with in-house Trehalose, with good aerosol performance mannitol, trehalose 32.0; compared to trehalose and or lactose as a Lactose, lactose which are moisture cryoprotectant 27.8 (< 5 sensitive µm) 27

Liposome with mannitol was not toxic to lung epithelial or alveolar macrophages, and did not activate inflammatory response in alveolar macrophages

Spray drying Mannitol, Rifampicin-PLGA Jethaler 28.3 7 Single step process to produce [77, (inlet, 60°C; ethylcellulose and submicron particle PLGA nanoparticles 78] outlet, 35-39°C; PLGA, either alone dispersed in mannitol fd, 2.5 mL/min; or in combination microparticles Rifampicin-PLGA recognised atom, 20 L/min) with by alveolar macrophages that polyethylenimine increased the rifampicin uptake

Freeze drying PLGA Single emulsion (w/o)- a Model 30 52 (1-5 µm) Non-biodegradable polyvinyl [79] solvent evaporation DP-4 dry alcohol replaced with powder degradable sucrose palmitate. insufflatorTM The particles efficiently (Penn phagocytosed by alveolar Century, macrophages PA)

Spray drying PLA and PLGA Polymeric microparticles Cyclohaler 28.3 PLA, 55- PLA showed faster release than [80] (inlet, 50°C) (1.1-5.8 70; PLGA. The microspheres µm) PLGA, 22- showed lack of cytotoxicity 32 profile and prolonged drug release in animal pharmacokinetic study

Spray drying PLGA Polymeric microparticles n/a n/a n/a Spray drying superior to single [81, (inlet, 60°C; emulsion to obtain high drug 82] outlet, 40°C; fd, payload particles compared to 16.7 mL/min; single emulsion technique asp, 15; atom, 700 Nl/h) Guinea pigs insufflated with single emulsion rifampicin- PLGA significantly reduced the bacterial load, reduced inflammation and lung damage compared to untreated group

Supercritical PLA Polymeric microparticles n/a n/a n/a Inhalable particles can be [83] precipitation produced using supercritical (CO2 flow, 1.5 technique g/mL)

Spray drying Stearic acid and Cationic lipopolymeric Rotahaler 28.3 67.9 (1.1– Process parameter optimisation [84] (inlet, 40°C; branched nanomicelle with Inhalac 5.8 μm) outlet, 33°C; fd, polyethyleneimine 230 as a carrier 2.5 mL/min; blended with Inhalac asp, 40 Nm3/h; 230 as a carrier atom, 0.8 kg/cm2)

Spray drying Leucine Large porous particles n/a n/a 53 (< 5.8 Pulmonary delivery using one- [85] (inlet, 170°C; µm) half of the oral dose achieved fd, 45 mL/min; systemic concentration similar asp, 25 g/min; to oral dose and 3-4-fold atom, 33 psi) higher drug concentration in bronchoalveolar lavage than the oral delivery in guinea pigs

Spray drying Chitosan particles Chitosan formulation with Lupihaler 28.3 10- 25 % The rifampicin loaded particles [86] (inlet, 150°C; blended with Lactohale® 100 and (<5 µm) were less toxic to the lungs outlet, 65°C; fd, inhalable lactose Lactohale 300® than the free drug treatment in 10 rpm; asp, 40 rats %; atom, 1.5 kg/m3)

Freeze drying Thiobarbituric acid Uniform and amorphous - - - The particles were mono- [87] rifampicin microspheres dispersed between 1 and 3 µm, and readily phagocytosed by macrophages Rifapentine Recrystallisation None Elongated crystalline Aerolizer 100 83.2 (< 5 The powder well tolerated by [30, and spray drying µm) alveolar macrophage and 88] (inlet, 60°C; epithelial cell lines outlet, 39°C; fd, 2 mL/min; asp, High drug concentration in the 40 m3/h; atom, lung compared to 650 L/h) intraperitoneal delivery in mouse model

Spray drying Hydrogenated soya Smooth, spherical Rotahaler 60 92.5 (< 4.4 Longer retention of drug in the [89, (inlet, 115- phosphatidylcholine particles of proliposomes µm) lungs of albino rats pulmonary 90] 125°C; outlet, pharmacokinetic 60-65°C; fd, 1.2 mL/min; asp, 35 Rifapentine loaded liposomes %; atom, 1.8 safe at 1 and 5 mg/kg. Higher bar) dose (10 mg/kg) caused liver and kidney damage in rat model

Spray drying PLGA Polymeric particles Osmohaler 100 52-57 (< 5 The molecular weight and [91] (inlet, 60°C; µm) monomer ratio is not crucial outlet, 39°C; fd, factors determining the aerosol 15 mL/min; asp, performance of PLGA particles 40 m3/h; atom, 672 L/h) 30

Rifabutin Spray drying Chitosan particles Chitosan formulation with Lupihaler 28.3 10- 25 % The rifabutin loaded particles [86] (inlet, 150°C; blended with Lactohale® 100 and (<5 µm) were less toxic to the lungs outlet, 65°C; fd, inhalable lactose Lactohale 300® than the free drug treatment in 10 rpm; asp, 40 rats %; atom, 1.5 kg/m3)

Isoniazid, Spray drying None Spherical with a surface Aerolizer 100 45 (< 5 µm) Combination powder displayed [92] rifampicin, (inlet, 60°C; composed primarily of better physical stability than and outlet, 39°C; fd, rifampicin the antibiotics spray dried pyrazinamide 2 mL/min; asp, alone 40 m3/h; atom, 500 L/h)

Isoniazid, Freeze drying Sucrose as a Thin film hydration n/a n/a n/a Rapid high drug concentration [93] rifampicin, cryoprotectant and method and liposomal in the lungs and the high drug and lactose core carrier nanoparticles coated with concentration retained over pyrazinamide mannan as ligand long period (nearly 24 h) when compared to free drug treatment

Isoniazid and Spray drying PLA Polymeric particles Animal 20 78.91 (< 4.6 Drug concentration in the lungs [94, rifabutin (inlet, 53-55°C; nose-only μm) 20 times higher in 95] outlet, 32-34°C; exposure microparticles inhaled animals fd, 4.5 mL/min; chamber compared to drug solution atom, 700-800 administered mice group Nl/H) Prolonged half-life of the drugs after single inhalation in rhesus macaques

Moxifloxacin, Spray drying With or without Spherical and largely Aerolizer 100 55.5 ± 1.9% Leucine protects the particles [96] rifapentine, (inlet, 60°C; leucine amorphous (without from moisture to prevent and outlet, 39°C; fd, leucine) and recrystallisation of pyrazinamide 1.4 mL/min; 63.6 ± 2.0% pyrazinamide asp, 40 m3/h; (with atom, 540 L/h) leucine)

Rifapentine Recrystallisation Leucine Elongated rifapentine and Osmohaler 100 72-75 % Verapamil enhanced the killing [97] and verapamil and spray drying corrugated verapamil (<5 µm) of M. tuberculosis within the (inlet, 60°C; particles infected macrophages. The outlet, 39°C; fd, powder stable up to 3 months 2 mL/min; asp, stored in 0 % relative humidity 40 m3/h; atom, 650 L/h)

n/a: not available; inlet: inlet temperature; outlet: outlet temperature; fd: feed rate; asp: aspirator; atom: atomiser; CO2 flow: carbon dioxide flow rate

Table 2: Summary of inhaled powder vaccine formulations with the processing parameters and the key findings.

Drug Production Excipient Formulation Inhaler Flow FPF Key findings Refs method description device rate (%) (L/min) Ad35-vectored Spray drying Mannitol, trehalose, Mannitol n/a n/a n/a Trehalose with dextran formed [98] TB AERAS- (inlet, 65- leucine, sucrose, based dry dispersible particles and appeared to be 402 125°C; outlet, cyclodextrin, dextran, powder good vaccine stabiliser 34-50°C; fd, inositol, histidine, 4.5 mL/min; polyvinylpyrrolidone, asp, 439–538 and tween 80 m3/h L/h; atom, 35 m3/h)

Bacillus Spray drying Leucine Bacteria in a A hand-held 28.3 n/a Spray drying superior to freeze drying [99, Calmette– (inlet, 100- matrix of dry powder due to former reduced loss of viable 100] Guérin (BCG) 125°C; outlet, leucine inhaler and bacteria 40°C; fd, 7.0 newborn mL/min; asp, inhaler Inhaled powder displayed enhanced 35 L/h) and device immunogenicity compared to freeze drying equivalent dose of subcutaneous administration in guinea pigs

Culp 1-6 and Spray drying Mannitol Mannitol with n/a n/a n/a Intact protein concentration after spray [39] MPT 83 (inlet, 50°C; 1 % protein dried. Intratracheal insufflations in conjugated to outlet, < 35°C; mice showed a significant influx of Lipokel fd, 1.0 neutrophils and antigen presenting mL/min; asp, cells in the lungs 38 m3/h; atom, 700-800 L/h)

Mtb Antigen Spray drying PLGA Polymeric Insufflator 60 68.9 PLGA-rAg85B induced hybridoma [101, 85B (Ag85B) (inlet, 65°C; microparticles DP-3 (Penn (< cells to produce IL-2 at intensity that 102] and trehalose outlet, 41- Century, 5 μm) was two orders higher than response by dibehenate 43°C; fd, 4.5 PA) free rAg85B (TDB) mL/min; asp, 600 L/h; atom, PLGA-rAg85B effective in boosting 3.0 bar) the primary BCG immunisation against M. tuberculosis in guinea pigs

n/a: not available; inlet: inlet temperature; outlet: outlet temperature; fd: feed rate; asp: aspirator; atom: atomiser

2.1 Formulations by spray drying

Spray drying is a widely explored technique for the production of respirable particles owing to its advantages of cost-effectiveness and ready capacity for scale-up. It has been successfully used to develop commercially available inhaled medications, such as insulin (Exubera®,

Pfizer), tobramycin (Tobi® Podhaler, Novartis) and mannitol (Aridol®, Pharmaxis) [103-105].

Spray drying is a single step process that directly converts drug solutions or suspensions into particles. Details of the spray drying process and its mechanism were reviewed by Vehring

[106]. Briefly, a drug solution or suspension is atomised into a drying chamber where the sprayed droplets are rapidly dried into solid particles by a heated gas. The size distribution and physical properties of the particles can be optimised through modifying process parameters such as the feed rate, atomisation pressure, inlet air temperature and aspirator gas flow.

Spray dried drug particles often exist in an amorphous state, which is usually cohesive, hygroscopic and has higher surface energy than the crystalline form. A conventional approach to improve aerosolisation of cohesive powders is spray drying with surface coating materials to reduce inter-particulate forces [106, 107]. Amino acids especially leucine is a well- recognised dispersion enhancer in spray dried powders [106, 108, 109]. Leucine is able to recrystallise on the surface of the particles, forming a shell to reduce inter-particle interaction and enhance the flow and dispersibility of the powder [108, 110]. Chan et al. reported that leucine could prevent water-induced recrystallisation of hygroscopic pyrazinamide in an inhalable formulation containing rifapentine, moxifloxacin and pyrazinamide [96]. Due to the low water solubility of leucine (24 g/L), this surface coating was also shown to delay the dissolution rate of the drug particles [111]. In the context of TB treatment, the slow release may prolong the drug action in the lungs and enhance the internalisation by macrophages. In

addition, surface coating with lipophilic excipients, such as magnesium stearate and sodium stearate, was demonstrated to reduce interparticulate contact area and deagglomeration for enhanced aerosol performance [112, 113].

Over the last decade, inhalable dry powder formulations consisting of first-line, second-line or a combination of both classes of anti-TB drugs have been studied to reflect current treatment regimens for the treatment of MDR and XDR cases [92, 94, 96, 114, 115]. There are also few studies reported on spray dried BCG and novel vaccine candidates [39, 99, 116]. Some of the well-established spray dried drug or vaccine formulations will be discussed.

2.1.1 Capreomycin sulphate

Capreomycin sulphate is a polypeptide antibiotic and a first choice of the second-line drugs in

MDR-TB chemotherapy. Capreomycin is administered intramuscularly or intravenously up to

1.0 g daily for 5 days a week [117] and these routes of administration showed a low peak serum drug concentration [118]. Garcia et al. reported pioneering study on inhaled capreomycin, showing a thin-walled porous particles produced by spray drying of 80 % capreomycin with

20 % leucine in ethanol:water cosolvent system [52]. The particles had a mass median aerodynamic diameter (MMAD) of about 5 µm with a reasonably high fine particle fraction

(FPF < 5.8 µm) of 47.7 % when tested at a flow rate of 28.3 L/min using a hand-held dry powder inhaler [52]. A single-dose insufflation of the formulation into guinea pigs resulted in a low serum drug concentration, but a longer half-life compared to intramuscular or intravenous administrations [52]. No significant changes in drug content, physical diameter and aerosol performance was observed after storing the powder at 4°C or room temperature and 10.5 % relative humidity (RH) [51]. In addition, the study investigated the influence of leucine

concentration on the aerosol performance of capreomycin. No significant difference were observed in the FPF (< 5.8 µm) of the formulations with 10 to 50 % leucine, suggesting 10 % was sufficient to improve powder dispersion [51]. Interestingly, denser particles were reported when leucine concentration was reduced from 50 to 30 % (0.36 g/mL for 50 % leucine; 0.68 g/mL for 30 % leucine), but the particles became less dense when the leucine concentration was further reduced from 20 to 10 % (0.64 g/mL for 20 % leucine; 0.32 g/mL for 10 % leucine)

[51].

More recently, Shoubben et al. produced a capreomycin formulation using a nano-spray dryer

[54]. The initial powder was cohesive, resulting in a low respirable fraction of only 14 % at 60

L/min using HandiHaler®. When the powder was blended with lactose (1:50 capreomycin:lactose) the respirable fraction was increased to 26 %, highlighting the need of lactose to improve the dispersability of the capreomycin. As mentioned earlier, the bulky composition of lactose carrier may not be practical for delivering a large dose of capreomycin required for TB therapy. In regards to the differences between conventional and nano spray dryer, the latter has an advantage of achieving high powder recovery with low amount of materials or/and low volumes of samples to allow formulation optimisation during early product development [119].

2.1.2 Rifamycins (rifampicin and rifapentine)

Rifampicin is a semi-synthetic derivative of rifamycin and the most potent sterilising agent in the current anti-TB regimen. The drug is orally administered at a recommended dosage of 600 mg daily [120]. Rifapentine is a long-acting derivative of rifampicin that was subsequently approved by Food and Drug Administration (FDA). Rifapentine may be administered weekly

at an oral dose of 600 mg during the continuation period of TB therapy in low risk and HIV- negative TB patients [23].

Rifampicin

Son and McConville produced a spray dried, flake-like rifampicin dihydrate for inhalation [69].

The rifampicin crystalline powder possessed superior dispersion properties with FPF (< 5 µm) above 60 % when dispersed using both Aerolizer® and HandiHaler® devices at 60 L/min. This formulation was chemically stable for up to nine months at ambient conditions, compared to approximately 26 % degradation in amorphous rifampicin under the same conditions [69]. In a subsequent study, the dihydrate crystals were coated with poly (lactic-co-glycolic acid)

(PLGA) and polylactic acid (PLA) polymers using a three-fluid spray nozzle to reduce the initial drug release, and this was shown to protect the microcrystals from water absorption [70].

However, this coating adversely affected the aerosol performance of the crystals as the FPF was reduced to less than 45 %, probably owing to the larger particle size after coating [70].

Ohashi et al. produced PLGA-rifampicin nanocomposite particles dispersed in mannitol microspheres without any stabiliser using a four-fluid nozzle spray dryer, as reviewed in Ozeki and Tagami [121]. Briefly, rifampicin/PLGA in acetone:methanol (2:1), aqueous mannitol and compressed air were simultaneously supplied through different passages of the four-fluid nozzle, mixed at a high turbulence at the nozzle tip and then spray dried [122]. In vitro studies in rat macrophages showed that these rifampicin nanoparticles were rapidly phagocytosed by alveolar macrophages leading to increased drug uptake [78]. These studies confirmed that it is possible to produce polymeric nanocomposites for efficient delivery of rifampicin to the lungs

using the four-fluid nozzle spray dryer in a single step process. This is potentially a rapid and readily scalable method to produce inhalable nanocomposite particles.

Rifapentine

When compared with rifampicin, rifapentine has the advantages of a longer serum half-life, greater bactericidal activity and higher intracellular accumulation [88, 123]. Chan et al. compared dispersibility of solvent spray dried amorphous and crystalline form of rifapentine produced through acetone-water cosolvent system [88]. Both forms were suitable for inhalation with a FPF (< 5 µm) of more than 70 % when dispersed at 100 L/min using the Aerolizer®.

The crystalline form, however, exhibited a greater chemical stability than the amorphous rifapentine after storage at both 0 % and 60 % RH for at least 3 months. In another study, Patil-

Gadhe et al. produced inhalable hydrogenated soya phosphatidylcholine (HPSC) proliposomes loaded with rifapentine using a single-step spray drying of hydroalcoholic solution containing rifapentine, HPSC, and cholesterol. The proliposomes have a low MMAD (1.56 ± 0.16 µm) and high FPF (< 4.4 µm, 92.5 ± 1.5 %) [89]. The powder also showed a good tolerability at doses up to 5 mg/kg with improved pulmokinetic parameters in rats [89, 90].

2.1.3 Para-aminosalicylic acid (PAS)

There has been growing interest in formulating para-aminosalicylic acid (PAS) into an inhalable form, in part due to the requirement of relatively large dose (8 - 12 g) when administered orally. Tsapis et al. formulated porous and partly crystalline micron-sized particles by spray drying PAS with 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) in ethanol solution [66]. These particles were suitable for optimal lung deposition (FPF < 5.6 µm,

59 - 63 % at 60 L/min using AIR® inhaler) and were physically stable up to four weeks at 40°C and 15 % RH [66]. On the other hand, Gad et al. produced nanoporous microparticles by spray drying with ammonium carbonate that underwent thermal decomposition in the hot drying chamber into gaseous ammonia, carbon dioxide and water thus producing a novel salt/complex form of PAS [65]. The particles were crystalline and agglomerated into non-uniform rough surface sphericals. When the powder was dispersed using the Spinhaler® device at 60 L/min, approximately 25 % drug deposited in the lower region of the impinger (< 6.4 µm) [65]. The

FPF of this pure drug formulation will need to be improved further to permit greater lung exposure to PAS with a lower inhalable dose than the typical oral dose.

2.1.4 Pretomanid (PA-824)

Pretomanid (PA-824) is a nitroimidazopyran anti-TB drug, currently in Phase II clinical trial as part of a combination with standard oral regimen for drug sensitive TB [124]. Sung et al. produced thin-walled, porous particles composed of PA-824, leucine and phospholipid 1,2- dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) for inhalation using spray drying technique

[63]. The MMAD, geometrical standard deviation (GSD), and FPF (< 5.8 µm) was 4.74 ± 0.04

µm, 1.53 ± 0.02 and 53.3 ± 2.7 %, respectively which support the suitability of the powder for inhalation [63]. The powder also exhibited intact physical, chemical and aerodynamic stability for a year under refrigerated condition (4°C), and for six months at room temperature (25°C) and 10 % RH [63]. Furthermore, the powder exhibited an elevated drug concentration in the lungs of the guinea pigs when given via the pulmonary route compared to the oral delivery

[63].

2.1.5 Isoniazid

There are various reports on inhalable isoniazid. Severe hepatotoxicity is an infrequent but important adverse effect of oral isoniazid, and direct drug delivery to the lungs is a rational approach to reduce this toxicity. Rojanarat et al. showed that spray dried isoniazid-loaded proliposomes possessed a relatively low FPF (< 4.4 µm) of 15 to 35 % when dispersed at 60

L/min using a locally-made dispersion device [59]. In contrast, an FPF (< 5 µm) of more than

60 % was reported for spray dried chitosan particles dispersed using Cyclohaler® at 28.3 L/min

[60]. The poor dispersion of proliposomes was due to strong agglomeration force between the particles compared to the less cohesive chitosan particles. In addition, Pourshahab et al. produced isoniazid-loaded chitosan/tripolyphosphate nanoparticles via an ionic gelation method and spray dried into microparticles using lactose and leucine as carriers [61]. The in vitro deposition data indicated the spray dried powder has a high FPF of 45 % at 60 L/min using Cyclohaler®. Furthermore, the nanoparticles showed enhanced in vitro anti- mycobacterial activity compared to isoniazid solution, and this was attributed to the nanoparticles binding to the negatively charged bacterial cell surface [61].

2.1.6 Combination formulations

For multi-drug formulations, the combination ratios were determined based on the recommended oral dose [92, 96]. Chan et al. produced an excipient-free dry powder consisting of pyrazinamide, rifampicin and isoniazid at a ratio of 5:2:1 (w/w) following the oral doses of

25 mg/kg, 10 mg/kg and 5 mg/kg, respectively [92]. The TOF-SIMS analysis of these particles showed that rifampicin predominantly concentrated on the surface of the particles. This observation suggested that during spray drying, the hydrophobic and heavier rifampicin

molecules tend to remain on the surface of the drying droplets to form the outer layer compared to the hydrophilic and smaller molecules of pyrazinamide and isoniazid. The combination formulation also displayed better physical stability than the spray dried individual drugs [92].

By contrast, previous studies reported that when rifampicin was combined with either isoniazid, pyrazinamide or both, it resulted in formation of adduct-rifamycin isonicotinyl hydrazone [125]. A similar phenomenon was also reported in polymeric formulations [126]. In the following study, Chan et al. replaced isoniazid and rifampicin with moxifloxacin and rifapentine, respectively, and spray dried with or without leucine as an excipient. The powder was chemically stable for up to 3 months storage at 25°C and 60 % RH and showed no extra peaks in the HPLC data [96]. The combination of spray dried drugs also retained their in vitro anti-mycobacterial activity. However, the stabilisation with leucine was necessary to prevent recrystallisation of pyrazinamide for long-term storage. Overall, these studies demonstrated the feasibility of spray drying drugs in combination into respirable powders without altering or degrading their chemical constitution or anti-mycobacterial properties.

2.1.7 Vaccine candidates

The general principle for spray drying vaccine formulations is similar to that for drugs, except that careful attention is required on the processing parameters, such as excipient selection and drying temperature, to prevent denaturation of antigens. These aspects of spray drying vaccines along with stabilisation process were systematically reviewed by McAdams et al. [127].

For inhalational delivery, live M. bovis BCG (5 % w/w) was formulated into an inhalable powder containing a larger fraction of L-leucine (95 % w/w) to improve the flowability and dispersibility of the powder [99, 100]. The powder showed intact bacterial viability upon

storage at 4°C and 20 % RH in high moisture protection vials up to 3 months [99]. The dry powder form of BCG delivered intra-tracheally to guinea pigs induced protection against pulmonary TB [100].

Recently, Tyne et al. successfully produced dry powders containing the M. tuberculosis antigens, cutinase-like proteins (Culp) 1 and 6 and MPT83, conjugated to Pam2Cys (Lipokel), a TLR2 ligand, with excipient mannitol [39]. Mannitol was chosen because unlike other sugars, such as sucrose and lactose, spray dried mannitol is largely crystalline and physically stable

[128]. Immunoblotting revealed that the protein-adjuvant conjugates remained intact after spray drying [39]. Intra-tracheal delivery of both protein-adjuvant conjugates stimulated protective T cell responses in the lungs of the vaccinated C57BL/6 mice [39]. Using a similar approach, Jin et al. spray dried adenoviral-vectored vaccine (AERAS-402) in combination with mannitol, cyclodexrin, trehalose and dextran. The powder was reported to be hydrophobic which benefited the survival of the viruses vaccine during both production and storage at various temperatures covering 4 - 37°C [98]. These findings indicated dry powder vaccines would have a prolonged shelf life in the absence of cold-chain storage, and this would make the aerosol immunisation practical globally.

2.1.8 Bacteriophage formulations

Bacteriophage (or phage) therapy is being reappraised as an alternative treatment to antibiotic therapy [129, 130]. Mycobacteriophages are viruses that infect M. tuberculosis thereby killing the pathogenic bacteria. Phages subsequently lyse the host bacteria, releasing the progeny that can infect mycobacteria at other sites of infection allowing continuation of the phage cycle.

Phages are a promising substitute to antibiotic therapy because resistance to phage infection is

unlikely to develop [131]. Some early studies reported successful phage inhalation therapy for treating patients with other lung infections [132, 133]. Recently, Liu et al. reported the first systematic study on aerosol delivery of mycobacteriophages D29 to mice utilising nebulisation mode [134]. Inhalable dry powder containing mycobacteriophages have not been formulated for TB therapy, however the optimal excipients and the stability of phage formulations have been investigated for other lung infections, such as in cystic fibrosis.

Excipients, including polyethylene glycol, mannitol, lactose, leucine, and trehalose, are used for spray drying phages to act as bulking and protectant agents during the process. Trehalose is often used to minimise the loss of phages during the processing, and leucine is utilised to improve the powder dispersibility in spray drying [135]. The mixture of trehalose, leucine and a surfactant can generate a spray dried phage formulation with low processing loss [136].

Vandenheuvel and colleagues tested the influence of different excipients, including lactose, trehalose and dextran 35 for spray drying the phages [137]. The results revealed that trehalose was the most efficient excipient for protecting the phages from mechanical stress and temperature during the spray drying process. In a study by Golshahi et al. phages were formulated with lactose and lactoferrin using freeze drying instead of spray drying [138].

Viability tests after freeze drying showed the loss of one and two log10 titer for phages KS4-M and PhiKZ, respectively. The phages also remained viable after a milling procedure for deagglomerating the freeze dried powders. Furthermore, Leung et al. compared the phage viability and dispersibility of combination powders containing mannitol, trehalose and leucine produced via spray drying and spray freeze drying. The spray dried powders caused less loss in phage titre and a higher respirable fraction than the spray freeze dried powders. However, the spray freeze dried particles were reported to provide a better physical protection to the phages than the spray dried counterparts [139].

To develop stable and inhalable formulation of phages, factors such as pH, temperature, interfacial adsorption and ionic strength, play a key role in the stability and the feasibility of the formulation. Knezeciv and colleagues tested the stability of Pseudomonas aeruginosa- specific phages in different pH conditions and demonstrated complete survival at pH 7. The survival rate dropped slightly to 80 - 100 % at pH 9; however, a significant drop was observed at pH 1.5 [140]. Studying the thermal stability of phages revealed their stability at 4°C.

Nonetheless, they become less stable at elevated temperatures, and this should be avoided in processing phage formulations [141]. Another important factor in phage formulation stability is ionic strength. Phages may undergo osmotic shock followed by death if there is a substantial difference in the osmotic pressure between the interior and exterior of the phage [138]. Hence, care should be taken during the drying process in order to reduce the rate of ionic strength change. Mechanical stress, such as high speed mixing, centrifugation, and filtration may damage the phages causing them to denture [142].

Matinkhoo et al. formulated inhalable phage powder with trehalose and leucine using a relatively low spray dryer outlet temperature of 40 to 45°C, and this resulted in less than one log10 titer loss after spray drying [136]. The thermal and shear stress during these processing parameters did not significantly affect the stability of the phages. By contrast, Finlay’s group tested the stability of freeze dried phage formulations designed for inhalation and found no significant titer loss when the phage powders were stored at 22 % RH at 4°C and 21°C for three months [138]. A recent study demonstrated the stability of spray dried phages with trehalose at low temperatures (4°C) and RH (0 %). However, they become increasingly unstable with rise in humidity owing to matrix recrystallisation that can damage the phages. In addition, some phages become unstable when stored at the relatively high temperature of 25°C [143]. In conclusion, formulations and storage conditions need to be optimised for each type of phage to

ensure stability. More work is required for examining the long-term stability of phage formulations in respirable forms.

2.2 Formulations by freeze drying and spray freeze drying

Freeze drying or lyophilisation is a process of drying frozen material through sublimation and desorption under vacuum. This method is ideal for heat sensitive samples including immunogenic proteins and peptides, and bacteriophages. The freeze drying of bacteriophages

KS4-M and PhiKZ with lactose:lactoferin (60:40, w/w) produced a readily dispersible powder that did not require milling. The powder demonstrated an acceptable FPF (< 5µm) of 32 – 34

% when dispersed using an Aerolizer® at 60 L/min [138]. Furthermore, the freeze drying technique has often been used for drying liposomal or polymeric formulations [56, 74], and this will be discussed in the following sections. For instance, Diab and colleagues produced rifampicin-loaded PLGA particles using single-emulsion (oil/water) method, followed by freeze drying [79]. Dispersion analysis using a powder insufflator (Model DP-4, Penn Century,

PA) at 30 L/min showed a high FPF, with approximately 52 % of the powder having aerodynamic diameters between 1 - 5 µm.

Another emerging technology in particle engineering that combines spray freezing and freeze drying steps is spray freeze drying. The drug solution or suspension is atomised into a cold medium (e.g. liquid nitrogen) to snap freeze the droplet, which are then dried under vacuum in a freeze dryer to produce large, light and porous particles with enhanced aerosol performance.

The production yield is almost 100 % [144, 145]. The cost of spray freeze drying can be high compared with other conventional processes due to the extensive use of liquid nitrogen.

Alternatively, an atmospheric spray freeze drying technology is available which may reduce the production cost [146].

2.3 Formulations by supercritical fluid precipitation

Supercritical fluid precipitation is efficient in producing uniform particles in term of size, crystallinity, morphology, and distribution to benefit dispersibility of the powder [147]. The technique is based on extraction of solvents from a supercritical fluid containing a dissolved compound [148]. A detailed review on application of this technology for pulmonary drug delivery is available in Sun [149].

Manion proposed a combination of supercritical fluid precipitation and spray drying to produce inhalable capreomycin, kanamycin and isoniazid particles [53]. The particles were corrugated spheres and largely amorphous with less than 3 % moisture content. These powders exhibited a respirable fraction (< 3.3 µm) of 21.6 % determined via the Puffhaler® at a flow rate of 28.3

L/min. Importantly, the processed formulations of these drugs as well as other anti-TB drugs, such as rifampicin, moxifloxacin and ciprofloxacin, retained in vitro anti-M. tuberculosis activity, confirming that the technique does not affect the anti-mycobacterial activities of the drugs [53, 150]. Alternatively, Ober et al. manufactured rifampicin/lactose microparticle composites using supercritical antisolvent drug-excipient mixing technique. Fourier transform infrared spectroscopy analysis demonstrated no chemical interaction between rifampicin and lactose [68]. The potential drawbacks of the technique include the requirement for high pressure and complicated scale-up procedures.

3. OTHER INHALABLE DRY POWDER PRODUCTION APPROACHES

3.1 Liposomal formulations

Liposomes are small spherical shaped vesicles created from a variety of non-toxic phospholipids. First described in the mid-60s [151, 152], the advantages of liposomes include the ability to achieve narrow particle size distribution and a high drug loading [153, 154].

Furthermore, encapsulating poor water soluble drug in liposomes resolves the solubility issues and negates the need for potentially toxic solubilising agents [155]. The slow release property of the liposomes also prolongs the drug residence in the lung and facilitates uptake by macrophages [156, 157]. For drugs that have an unpleasant taste or cause cough after inhalation, liposomes can effectively reduce these effects [155].

Among liposome formulations, distearyl phosphatidylcholine (DSPC), was the most suitable phospholipid to ensure vesicle stability and minimise lipid–drug interactions [153]. These liposomes are also suitable to encapsulate two or more drugs in a single liposomal formulation because of their hydrophobic and hydrophilic characteristics. For instance, Gursoy and colleagues successfully encapsulated hydrophilic isoniazid and hydrophobic rifampicin into a single formulation [158]. They even showed that co-encapsulation augmented the entrapment of the individual drugs.

The latest development in this field is cell-specific targeting in which the liposomes are coated with ligands to facilitate targeting of specific cells, such as alveolar macrophages [159].

Intermittent delivery of rifampin encapsulated liposomes grafted with macrophage activator tetrapeptide tuftsin was 2,000-fold more effective than the free drug in lowering the bacterial

burden in the lungs of M. tuberculosis-infected mice [160]. However, this study used liquid liposomal formulations that are typically unstable for long-term storage and require nebulisation for inhaled therapy [161]. The nebulisation process may cause significant loss of encapsulated drug and alteration in the physical integrity of the liposomes [155]. To overcome the instability and delivery constraints, liposomes can be dried using spray drying, freeze drying or spray freeze drying into proliposomes [144]. Proliposomes are free flowing particles that immediately form liposomes when in contact with aqueous media [162]. Rojanarat et al. developed levofloxacin and pyrazinamide proliposomes using spray drying [62, 67]. The aerodynamic analysis using an in-house dispersion device at 60 L/min showed an FPF (< 4.4

µm) and MMAD of approximately 30 % and 4 µm, respectively. Furthermore, repeated intratracheal treatment with the levofloxacin proliposome powder in rats showed no significant renal and liver toxicity. On the other hand, Shah and Misra prepared amikacin liposomes as a dry powder by freeze drying and blended them with lactose coarse carrier in varying mass ratios with or without fines to evaluate the influence of carrier on in vitro deposition performance [49]. Delivery with the Rotahaler® at a flow rate of 60 L/min showed that a lactose carrier containing 10 % fines (w/w) provided the optimum blend at 1:5 mass ratio of liposome:lactose having a FPF of 29 % [49]. Alternatively, Changsan and colleagues encapsulated rifampicin into liposomes and freeze dried these either with mannitol, trehalose or lactose as a cryoprotectant [75, 76]. Among the excipients tested, mannitol was the most suitable cryoprotectant, and it was crystalline after being freeze dried. The aerosol performance evaluation using an in-house dry powder inhaler at 60 L/min displayed a high FPF of 66.8 %

(< 5 µm) with a low MMAD of 3.4 µm, which is suitable for inhalation. The cytotoxicity assays showed rifampicin-loaded liposomes were nontoxic to bronchial epithelial cells, small airway epithelial cells, and alveolar macrophages. Rifampicin in the liposomal dry powder formulation was also found to be more stable compared to rifampicin aqueous solution and rifampicin

liposomal suspension.

Furthermore, Bhardwaj et al. successfully produced freeze dried isoniazid, rifampicin and pyrazinamide sub-micron liposomes using sucrose as the matrix. Scanning electron microscopy demonstrated that the liposomes adhered to the sucrose particles. The flowability of the powder was expressed by angle of repose that ranged between 18 - 20˚, and this was considered excellent for inhalation [93, 163]. Although these results demonstrate the suitability of liposomal vesicles for pulmonary delivery of anti-TB drugs, the high production cost and complex processing provide potential barriers to large scale production.

3.2 Polymeric formulations

Polymer-based carriers offer numerous advantages for inhalation delivery because of their ability to increase the solubility of hydrophobic compounds, the sustained release of the incorporated drugs, the potential to modifiy particle surface properties and protect the incorporated drugs from degradation [164]. As mentioned above, polymeric particles also increase uptake by macrophages. Hydrophobic PLA and PLGA have been among the most attractive candidates studied for inhalation delivery, owing to their biocompatibility and biodegradability and extended release profile. Further, they are FDA-approved polymers for systemic delivery [165, 166] but these are yet to be approved for inhalation. Therefore pulmonary delivery systems based on these polymers are of experimental nature. The accumulation of polymeric materials in the lungs over long-term inhalation use and the acidity of degradation products of these polymers may be of safety concerns.

Traditional methods of producing drugs encapsulated in PLA and PLGA particles utilised single (oil-water) or double (water-oil-water) emulsion techniques. In the single-emulsion method, a hydrophobic drug and polymer are dissolved in an organic phase (oil phase), and this solution is emulsified with a surfactant (water phase). By contrast in the double-emulsion method, a hydrophilic drug solution (water phase) would first be emulsified in an organic solvent containing a hydrophobic drug and polymer (oil phase), followed by a second emulsification in a surfactant (water phase). The resulting suspensions are stirred for several hours to allow evaporation of the organic solvent and hardening of the particles. The hardened particles are then washed and dried through spray drying, freeze drying or spray freeze drying techniques. Muttil and colleagues encapsulated two anti-TB drugs (hydrophilic isoniazid and hydrophobic rifabutin) within PLA microspheres through double-emulsion and spray drying

[94]. The powder showed a high FPF (< 4.6 µm) of 78.91 % when dispersed at 20 L/min into an animal nose-only exposure chamber for delivering drugs to mice via pulmonary route [94].

Using a similar approach, Lu et al. produced inhalable PLGA powder encapsulating M. tuberculosis antigen 85B (Ag85B) with the adjuvant trehalose dibenate (TDB). The dispersion analysis was performed using a Insufflator DP-3 (Penn Century, PA) at a flow rate of 60 L/min and showed a respirable fraction of 68.9 % (< 5 µm) [101].

Size distribution and drug loading of emulsified polymeric particles are highly dependent on the shear forces and formulation conditions, such as surfactant concentration, polymer viscosity and concentration, and the volume of the aqueous phase. Studies have suggested more advanced methods to achieve high drug loading with monodisperse particles, such as pre-mixed membrane emulsification and supercritical anti-solvent precipitation [73, 74, 83]. However, these technologies are not practicable for bulk production as they are expensive and labor intensive. Alternatively, a direct spray drying of the organic solvent containing polymer and

drug is a rapid and less laborious approach to achieve a high drug loading [80, 81, 91]. A 100

% drug loading efficiency can be attained from this technique because the drug cannot partition into the secondary water phase. Coowanitwang and colleagues produced inhalable rifampicin

PLA and PLGA powders using a spray drying and reported MMAD of 4.65 - 5.11 µm and 2.22

- 2.86 µm, respectively, upon dispersion using a Cyclohaler® at 28.3 L/min [80]. O’Hara and

Hickey, however, showed that direct solution spray drying resulted in higher concentrations of rifampicin on the surface of the PLGA particles [81], and this caused a greater initial release of drug compared to single emulsion particles.

Other polymeric formulation techniques include iontrophic gelation for preparation of tripolyphosphate-crosslinked chitosan microspheres with approximately 50 % isoniazid loading and an FPF of 67 % using Cyclohaler® at 28.3 L/min [60]. In another study, Pai et al. showed intratracheal instillation of rifampicin and rifapentine loaded chitosan particles suspended in phosphate buffer saline (PBS, pH 7.4) only caused mild changes in the lung pathology, compared to the severe peribronchiolar infiltration of inflammatory cells caused by free drug treatment in rats [86]. These findings suggest chitosan polymer dry powder is a potential platform for pulmonary treatment of TB.

Polymeric nanoparticles have also attracted considerable interests as drug carriers for pulmonary drug delivery systems. Nano-sized particles are readily internalised by alveolar macrophages and provide a sustained release reservoir in the vicinity of the target [167-172].

However when inhaled as a dry powder, nanoparticles can be readily exhaled because of the small aerodynamic diameter of the particles. They are also nearly impossible to deagglomerate and will aggregate quickly upon inhalation due to van der Waal forces [173]. One approach to overcome this problem is to blend the nanoparticles with coarse carriers. For example,

Vadakkan et al. produced spray dried spherical rifampicin cationic lipopolymeric nanomicelles and blended them with inhalation grade lactose (Inhalac 230). It was found that the spherical nanomicelles were adsorbed on the asperities of the lactose crystals that acted as “active sides”.

The powder was also in an inhalable range with a FPF (1-5 µm) of 67.9 % and a MMAD of

2.31 - 2.45 µm, when aerosolised with a Rotahaler® at a flow rate of 28.3 L/min [84]. However, this formulation was too bulky for inhalation owing to the addition of lactose carrier.

Nonetheless, the findings suggest that dry powder nanoparticles for pulmonary treatment of

TB are feasible.

4. DRY POWDER DELIVERY DEVICES

For effective anti-TB pulmonary regimen, patients may need to inhale a large dose of antibiotics. In Phase 1 clinical study, administration of an inhaled dose of 300 mg capreomycin

(12 capsules of 25 mg powder each) using a capsule-based handheld device resulted in a systemic concentration equal to or higher than the minimum inhibitory concentration (MIC) of the drug against M. tuberculosis [29]. These findings suggest that the patient may need to inhale repeatedly a large dose of drugs for an effective treatment outcome.

Both the nebuliser and dry powder inhaler (DPI) are capable of delivering large doses. The most common nebuliser is the conventional air-jet nebuliser that generates aerosol medication from liquid or suspension formulations using a compressed gas. Jet nebulisers are relatively cheap but bulky, noisy and cause wastage of medication owing to retention of drug liquid in the dead space of the device. Other newer air-jet nebulisers are breath enhanced devices, such as the Pari LC®, which minimise the drug wastage by mechanically controlling the delivery of the aerosol dose during the inspiratory phase [174]. A different type of nebulisers is based on

vibrating-mesh technology, which generates the aerosol by extruding the drug solution through a vibrating mesh containing an array of tiny pores. Examples are the Aeroneb® and Aerogen®, which are more handy to use and operate silently [175]. However, these nebulisers are relatively expensive and there is currently no data on dose adjustment to prevent overdose of medications that are already optimised for jet nebulisers [175, 176]. The DPIs are breath- activated devices for powder aerosol generation and delivery. The powder is dispersed into particles by the patient’s inspiratory flow and in general the device is inexpensive, hence suitable for application for TB [177]. Figure 1 depicts some of the currently available DPI products.

Currently, marketed DPIs that can deliver more than 100 mg of dry powder are the Novartis

TOBI Podhaler® (112 mg of tobramycin) and Pharmaxis Bronchitol® (400 mg of mannitol), both for the treatment of cystic fibrosis [104, 178]. One of the rate-limiting factors for these devices is that patients are required to load and inhale multiple capsules, and this may reduce patient adherence to the therapy, especially in the context of the lengthy anti-TB regimen [179,

180]. To overcome this limitation, we modified the Aerolizer® device to accommodate the size

0 capsule. Similar to Aerolizer®, the dispersion mechanisms of the modified device involves air turbulence and impaction inside the swirling chamber for deagglomeration of the powder

[181]. The performance of the device to deliver needle-like crystalline rifapentine powder in a single capsule (with loading of 30 mg) with multiple manoeuvres was tested at 60 L/min using multi-stage liquid impinger [88]. The dispersion analysis showed an FPF (< 5 µm) of more than 50 % with less than 10 % device and capsule retention, suggesting that the deagglomeration had occurred before the powder exited the device [182]. Traini and colleagues investigated the OrbitalTM, a breath-actuated (passive) and a single use disposable DPI [183].

This device consists of a sample compartment puck that releases the powder by centrifugal

force through one or more orifices into a deagglomeration chamber, where the powder breaks into inhalable size before delivery into the respiratory system. In vitro aerosol analysis reported that the device is capable of delivering a high quantity of mannitol co-spray dried ciprofloxacin or azithromycin formulations (ranging up to 400 mg) via multiple inhalation manoeuvres [183,

184]. For example, the mannitol-ciprofloxacin combination powder at a loading of 200 mg showed a FPF (< 5 µm) of 42.0 ± 1.8 % at 60 L/min, which was equivalent efficiency to the commercially available RS01 device (FPF < 5 µm, 38.4 ± 2.0 %) [183]. However, as the

OrbitalTM empties a relatively large amount (i.e. 100 mg) of powder with each actuation, it may cause side effects such as coughing or wheezing. Nevertheless, the potential of using the

OrbitalTM for TB therapy is being explored [185].

A device with a “open-inhale-close” routine would be preferred for inhaled drug delivery for

TB. In addition to important requirement of delivering a large dose, the device should have a simple operation process with flexible dosing to promote patient compliance. A possible example is TwinCaps®, a single-dose and disposable passive device used to deliver 40 mg of

Inavir® (laninamivir octanoate, a long-acting neuraminidase inhibitor) for the treatment of influenza [176, 186]. The device has been approved for children under 10 years old [176, 186].

TwincerTM is another single-use disposable passive device that consists of three plastic plates which can be simply stacked and clicked together. The device showed good dispersion and moisture protection of pure micronised colistin sulfomethate at doses up to 25 mg for the treatment of cyctic fibrosis [187]. Further optimisation of the TwincerTM (Cyclops) increased the loading dose of spray dried tobramycin to at least 50 mg [188]. Besides drug delivery, these devices are also applicable for pulmonary vaccination that requires a limited number of doses.

Most passive DPIs require a sufficient inspiratory flow rate (e.g. 30 - 60 L/min) to aerosolise powder from DPI, restricting their use for children (< 6 years old) and elderly patients [189,

190]. Therefore, active DPIs open a new window for pediatric and geriatric inhaled therapies.

Unlike passive devices, the dispersion of active devices is independent of inspiration flow.

Dispersion is usually achieved by utilising external energy sources such as: (1) the piezoelectric system to facilitate the aggregation of drug from dosing chamber (e.g. MicroDose Therapeutx

Dry Powder Inhaler); (2) embodiment with compressed air or spring to actively disperse the powder (e.g. Exubera® and Aspirair®); and (3) battery-driven motors to enable aerosolisation of powder (e.g. SpirosTM) [176, 186, 191-194]. Although these developments are proven to increase the lung deposition independent of the inspiratory flow, these alterations may reduce the portability and incur additional production costs.

TB patients may require admission to an intensive care unit (ICU) because of acute respiratory or multi-organ failure [195]. Unfortunately, DPIs are generally regarded as unsuitable for patients in ICU owing to their cumbersome administration. Recently, Tang et al. developed a novel dry powder delivery system for intubated patients that utilises a ventilation bag, a one- way duck-billed valve and an Osmohaler® [196]. The setup was capable of delivering Relenza®

(Zanamivir dry powder) with a reasonable FPF (< 5 µm, 13.6 %) in vitro [197]. The clinical safety of the system was also established using dry powder mannitol [198]. The findings support the potential of this strategy for inhaled therapy to intubated TB patients.

a b c

Figure 1. Photographs of some currently available dry powder inhaler devices; (a)

Osmohaler®, (b) HandiHaler®, and (c) Aerolizer®.

5. PRE-CLINICAL AND CLINICAL TRIAL EVALUATIONS OF INHALED TB

DRUG POWDERS

5.1 In vitro studies

Several in vitro cell culture models have been used to examine the anti-mycobacterial activity of inhaled drug formulations. Although primary alveolar macrophages are the most relevant cells for intracellular infection with M. tuberculosis, they are not readily accessible and may be heterogeneous. Therefore immortalised, macrophage cell lines, such as THP-1 cell line- derived macrophages are used to provide a homogeneous population that mimics the behavior of primary human alveolar macrophages [199]. Lawlor and colleagues used PLGA encapsulation to develop rifampicin, PAS and capreomycin dry powders that allowed a high respirable dose with Spinhaler® and Diskhaler® at a flow rate of 100 L/min [55]. Treatment of

THP-1-derived macrophages infected with M. tuberculosis H37Ra with the capreomycin and

PAS microparticles resulted in a significant mycobacterial killing. A more recent study showed that rifampicin-loaded freeze dried liposomes released the drug in a controlled manner, and the

uptake of encapsulated rifampicin into macrophages was more rapid than the free drug treatment [200].

The potential cytotoxicity of the dry powder drugs may also be determined using lung-derived cell lines, such as alveolar macrophages or lung epithelial cells. For example, a pyrazinamide- proliposome formulation was not toxic to human bronchial epithelial (Calu-3), human lung adenocarcinoma (A549) and rat alveolar macrophage (NR8383) cell lines. Further, the proliposomes did not activate macrophages to release inflammatory mediators in vitro [67].

An in vitro model using lung epithelial cells has proved useful to study the kinetics and mechanism of transepithelial transport of inhaled drugs. A detailed review on this model to assess the pulmonary absorption and disposition of drug particles is available in Sakagami

[201] and Steimer et al. [202]. Salomon et al. tested the permeability of CPZEN-45, a capramycin derivative with anti-TB activity, using Calu-3 monolayers. The permeability coefficients in absorptive and secretive directions were in the acceptable range of 0.43 ± 0.20

× 10-6 cm/s and 0.38 ± 0.12 × 10-6 cm/s, respectively [203].

5.2 In vivo studies

Several animal models have been established to study pulmonary TB as reviewed by

McMurray et al. [204]. Over the past five years, in vivo studies using mice [30, 58], guinea pigs

[52, 64, 205], rats [206], and monkeys [95] have investigated the efficacy of dry powder formulations for treatment of TB. Verma’s group used infection in Swiss mice to compare the efficacy of microparticles containing clofazimine delivered via the aerosol and oral routes.

Aerosol delivery reduced M. tuberculosis colony forming units (CFU) in the lungs by log10 2.6

compared to a reduction of log10 0.7 in the orally treated group [58]. In another study, crystalline rifapentine was administered by either intratracheal or intraperitoneal to BALB/c mice to study the pharmacokinetics of the drug [30]. The intra-tracheal delivery resulted in 38-fold higher concentrations of rifapentine in the lung than the plasma. Similar results were also reported using inhalable amorphous rifampicin in a rat model [206]. In addition, Garcia-Contreras et al. showed that guinea pigs receiving a high dose of inhaled PA-824 particles had lower degree of inflammation, tissue damage and bacterial burden after M. tuberculosis infection than the untreated or placebo-treated animals [64]. Aerosol delivery of BCG vaccine to guinea pigs subsequently challenged with virulent M. tuberculosis significantly reduced bacterial burden

(log10 3.2 reduction) and lung pathology compared to both untreated guinea pigs and animals immunised with standard parenteral BCG (log10 1.1 reduction) [21]. The major advantage of using mice, rat or guinea pig model is that there are similarities in the pathophysiology of TB with human and their immune system is well characterised. In addition, the purchase, husbandry and housing are less costly than for nonhuman primates.

Misra et al. investigated the biodistribution and pharmacokinetics of isoniazid and rifabutin delivered as PLA microparticles in monkeys via either inhalation or intravenous routes [95].

The elimination half-life of isoniazid and rifabutin were higher when the drug was inhaled.

Moreover, the drug concentration in the lung was twice that of the liver and four times that of the kidney. Previously, low dose infection of monkeys with M. tuberculosis was shown to closely resembled the pattern of disease in humans [207]. Despite the costly nature of the nonhuman primate infection, this model can be extremely beneficial in testing safety and efficacy of new formulations of drugs and vaccines, especially for late-stage preclinical studies.

Nonetheless, the in vivo studies consistently show that drug administration by inhalation has

the potential to increase selectively therapeutic efficacy at site of lung infection as well as to minimise systemic exposure and possible toxicity.

5.3 Clinical trial

Over the past ten years, only one paper has been published on a clinical trial using inhalation therapy targeting for pulmonary TB treatment. Dharmadhikari and colleagues conducted a

Phase I clinical study examining the safety and tolerability of capreomycin in 20 healthy subjects and measured the pharmacokinetic parameters [29]. Capreomycin was formulated by spray drying with leucine and administered at single doses of 25 mg, 75 mg, 150 mg, or 300 mg. Five subjects experienced temporary mild/moderate cough, but overall, the inhaled drug was well tolerated. No changes in pulmonary, renal and liver functions were detected post- inhalation. Rapid drug adsorption through the lung was detected as evidenced by the plasma concentration within 20 minutes. The inhalation of 300 mg of powder permitted a systemic

Cmax of 2.3 µg/ml compared with 32.0 µg/ml when 1.0 g of drug was administered intramuscularly [117]. However, a higher drug concentration was expected in the lungs after inhalation than after intramuscular delivery. The half-life (t1/2) of capreomycin was also extended to 4.8 h compared with 2 to 3 h after intramuscular administration [208]. These findings are significant, demonstrating that the inhaled capreomycin can achieve high drug concentrations at the infection site and longer drug retention in the systemic circulation, which would have the potential to reduce the drug dose and dosing frequency, and consequently drug- associated adverse effects. However, these inhaled pharmacokinetic data should be interpreted with caution as the studies were performed on healthy volunteers and they may not necessarily be replicable in TB patients. For example, in cystic fibrosis the increased bronchial circulation and pulmonary inflammation enhanced the absorption rate of inhaled drugs so that the dose-

normalised Cmax value of inhaled dry powder tobramycin was higher in cystic fibrosis patients than in healthy volunteers [209-211]. There is still a lack of understanding on respiratory physiology of TB patients, and this aspect clearly needs further clinical investigations to address the fate of particles deposited in the lungs.

The Clinical Trials Registry reports one ongoing clinical trial in Thailand on the safety of inhaled delivery of rifampicin, isoniazid, pyrazinamide and levofloxacin as individual dry powders [212].

6. RISK RELATED TO INHALED REGIMEN IN THE TREATMENT OF TB

Despite the potential benefits of inhaled therapy compared to the conventional regimen, there are also risks associated with direct deposition of particles in the lung. The lung has a complex biological defense system rich with immune cells, such as macrophages and leukocytes, and its vital role of gas exchange is dependent on the lipid-rich surface layer of surfactant and lung fluid homoeostasis. Any damage to these components could be life threatening. Clinical trials with large number of subjects are yet to be performed to determine the safety profiles of inhaled aerosol therapy for TB.

Besides the concern of drug toxicity, a potential risk of pulmonary drug delivery is inflammation of the respiratory tract caused by repeated deposition of inhaled particles. This may be more critical when treating pulmonary TB patients because of prior inflammatory changes in the lungs and the recruitment of mutiple immune cells, including monocytes, T and

B lymphocytes and natural killer cells [213, 214]. Furthermore, the deposition of foreign materials in the lungs may cause allergic inflammatory response and reduce lung function.

Misra and team demonstrated that intracellular delivery of inhalable microparticles loaded with anti-TB drugs activated M. tuberculosis infected macrophages [35, 215]. These included up- regulation of cytokines TNF-α, interleukin-12 (IL-12)), reactive nitrogen and oxygen radicals, and caspase-independent apoptosis [35, 215-217]. These products could potentially cause severe immunopathology through inflammation and oxidative stress [218]. These microparticles were intended for a slow drug release profile indicating they could stay in the lungs for some time [95]. Therefore, such a prolonged adverse immunological phenomenon may also lead to hapten-specific immunoglobulin E (IgE) responses and extensive mast-cell degranulation [218].

Unlike delivery of drugs, vaccines do not require multiple doses over a long period, but there are still potential risks of immunopathological reactions. Adjuvants do induce some inflammatory response as part of their immune stimulation, but excessive inflammation may be harmful. For example, Simecka and colleagues showed that intranasal immunisation with cholera toxin adjuvant caused a massive pulmonary airway inflammation characterised by IgE production and significant infiltration of mononuclear cells into the lungs, exclusively around pulmonary airways and vessels [219, 220]. This phenomenon resulted in bronchoconstriction of the airways and/or airway blockage with exudates leading to reduction in the airflow [220].

Similarly, pulmonary vaccination may induce significant immune response characterised by hypersensitivity reactions in the lungs which possibly contributing to pathologic changes in the lung. In addition, aberrant host innate immune responses characterised by dysregulation of cytokine or “cytokine storm” in the lungs can be dangerous [221]. The best example is the influenza pandemic of 1918 where the high degree of mortality was caused by the up-regulation of inflammatory chemokines and recruitment of neutrophils in the bronchi [222]. A similar scenario can occur when antigents directly deposited in the highly immunogenic lung region.

Rational approaches need to be developed to avoid serious immunogenic or inflammatory reactions associated with pulmonary delivery.

For drug delivery, a minimum proportion of a non-toxic excipient or carrier is critical to achieve desired therapeutic outcome with good aerosol performance [223]. A balance between maximum drug loading and carrier system is critical for inhalation product. The α-lactose monohydrate is the only FDA approved sugar that can be used as a carrier in dry powder formulation to enhance the flowability and dispersion of the powder [48]. However, as mentioned earlier, this approach is not very practical for TB because of the requirement of high doses. Other dispersion enhancer including 10 - 20 % leucine, is often used to balance the maximum drug loading and aerosol performance [51, 224]. Administration of a single dose of

300 mg capreomycin:leucine at 8:2, which is equivalent to 60 mg of leucine, showed good tolerability in healthy volunteers [29]. The safety of multiple deposition of leucine in the lungs is still unknown.

The inhalable vaccine formulations must stimulate immune responses without causing pathological damage. One successful adjuvant was the positively-charged chitosan particles that contained a DNA vaccine encoding HLA-A*0201-restricted T-epitopes of M. tuberculosis.

The chitosan acted as an adjuvant to promote the secretion of Interferon-γ by T cells recruited to the lungs following pulmonary immunisation without lung damage [36, 225]. By contrast,

PLA and PLGA microspheres induced pro-inflammatory responses following phagocytosis by macrophages [226] and these may limit the benefit of T cell responses induced by the encapsulated agents. Furthermore, drug containing PLA and PLGA particles were reported to provoke TNF-α, IL-12, and reactive oxygen and nitrogen radical production in macrophages

recovered from infected mice [215], while induction of apoptosis in infected THP-1 derived macrophages [216].

7. EXPERT OPINION ON FUTURE DEVELOPMENT

The promising results from various in vivo studies and the Phase 1 human clinical trial provide evidences for performing larger scale efficacy and safety inhaled anti-TB clinical trials.

Nonetheless, there are a number of challenges that need to be addressed.

1. Physiology of TB patients and particle deposition. There is a lack of understanding of lung and airways physiology of TB patients. The airflow of patients may differ from healthy population and this could affect the deposition of particles and their fate in the lungs. The intended region for powder deposition is also determined by the formulation and device designs. Tuberculous granulomas may be distributed throughout the lung and therefore more diffused deposition of the drug aerosol in the lungs is desirable. Thus, a greater understanding of the lung and airways of TB patients and the correlation with the aerosol deposition is crucial to facilitate successful inhaled anti-TB therapies [227].

2. Dosing consideration and strategies. Currently, prolonged therapy with a combination of anti-TB drugs for at least 6 months is essential for optimal clinical outcomes and to prevent the emergence of drug-resistant M. tuberculosis strains. Inhaled antibiotics may play a role as an adjunct to the existing oral therapy by providing high concentration of drugs at the site of infection [228]. Multiple dosing may be needed for an optimal pulmonary drug concentration, depending on the powder formulation and pharmacokinetic of the drug. The frequency of dosing and the long-term effects of pulmonary delivery remain to be investigated.

3. Screening of drug/vaccine candidates. Not all drugs/vaccines are ideal for inhaled therapy and there should be careful selection of the drugs/vaccines for testing by the pulmonary route.

These candidates should be highly efficacious and able to reduce the dosing frequency when given via inhalation compared to the conventional route of administration. Cheaper manufacturing and packaging procedures along with inexpensive inhaler devices will be advantageous.

4. Medication cost and feasibility. If the medication is expensive to be manufactured into an inhalable form, it will impact the global accessibility, especially to the resource-poor countries with a high burden of TB. Furthermore, a correct inhaler handling technique is important for an optimal benefit from inhaled therapeutics. This will require training for the TB clinic staff on proper usage of the inhaler to educate the patients, and this may incur additional costs. The lower development costs and ease of oral administration probably make many organisations including WHO, favour the discovery of new anti-TB drugs for oral administration instead of reformulations of approved anti-TB drug into an inhalable dosage form [227]. The Phase 1 trial of an inhaled pharmacokinetic study with capreomycin addressed the benefits and barriers of progressing the preclinical studies of inhalation therapies to practical treatment for TB [29].

Therefore cost-benefit and feasibility studies are necessary to justify the global implementation of inhaled therapy for TB [227].

8. CONCLUSIONS

Pulmonary administration using powder aerosols is a relatively new approach for the treatment of/or vaccination against, pulmonary TB especially targeting for MDR and XDR cases. An optimised powder formulation coupled to a suitable inhaler device is critical for effective pulmonary delivery and intended clinical outcomes. A substantial amount of powder formulation works has been undertaken over the past two decades aiming at delivering inhaled

TB therapeutics to the lungs. A common theme is to develop carrier-free formulations to increase the drug load required for high-dose antibiotics. They also avoid any carrier-associated toxicity. Combination drug formulations are necessary for effective treatment of TB and for reducing the emergence of drug resistance strains. Formulations of micro- or nano-particles with a slow release profile will promote phagocytosis by M. tuberculosis-infected alveolar macrophages. An inhaler device with an “open-inhale-close” routine will enhance patients’ compliance for the treatment, but both the generation and administration of a high powder dose remain a challenge to be tackled satisfactorily. It is also a prerequisite for the formulation/device to be cost-effective for countries with high burden of TB cases. While the formulation development and preclinical evaluation are in their advanced stage, clinical trials are significantly lagging behind, and this is partly due to the safety concerns of administering inhaled drugs or vaccines to the infected lungs that are already inflamed. However, if the biosafety hurdles can be mitigated and the TB drugs or vaccines efficiently delivered to TB patients, then inhaled TB therapy has potential to contribute to the global elimination of TB.

ACKNOWLEDGEMENT

The authors acknowledge the financial support from the Australian Research Council’s

Discovery Project funding scheme (DP120102778) and the National Health and Medical

Research Council’s Centre of Research Excellence in Tuberculosis Control (APP1043225).

Thaigarajan Parumasivam is a recipient of a Malaysian Government Scholarship.

REFERENCES

[1] World Health Organization, Global tuberculosis report 2014, in, http://www.who.int/tb/publications/global_report/gtbr14_main_text.pdf?ua=1, 2013. [2] E.L. Corbett, C.J. Watt, N. Walker, D. Maher, B.G. Williams, C. Dye, The growing burden of tuberculosis: global trends and interactions with the HIV epidemic, Arch Intern Med, 163 (2003) 1009-1021. [3] A. Zumla, A. George, V. Sharma, R.H.N. Herbert, A. Oxley, M. Oliver, The WHO 2014 global tuberculosis report—further to go, Lancet Glob Health, 3 (2015) e10-e12. [4] N.W. Schluger, The pathogenesis of tuberculosis: : the first one hundred (and twenty-three) years, Am J Respir Cell Mol Biol, 32 (2005) 251-256. [5] I. Smith, Mycobacterium tuberculosis pathogenesis and molecular determinants of virulence, Clin Microbiol Rev, 16 (2003) 463-496. [6] A. O’Garra, W. J. Britton, Cytokines in tuberculosis, in: S. H. E. Kaufmann, W. J. Britton (Eds.) Immunology and cell biology, Wiley-VCH, Weinheim, 2008, pp. 185-225. [7] J.L. Segovia-Juarez, S. Ganguli, D. Kirschner, Identifying control mechanisms of granuloma formation during M. tuberculosis infection using an agent-based model, J Theor Biol, 231 (2004) 357-376. [8] B. M. Saunders, W. J. Britton, Life and death in the granuloma: immunopathology of tuberculosis, Immunol Cell Biol, 85 (2007) 103-111. [9] D. R. Roach, A.G.D. Bean, C. Demangel, M.P. France, H. Briscoe, W.J. Britton, Tumor Necrosis Factor regulates chemokine induction essential for cell recruitment, granuloma formation and clearance of mycobacterial infection, J Immunol, 168 (2002) 4620-4627. [10] R.M. Jasmer, P. Nahid, P.C. Hopewell, Latent tuberculosis infection, NEJM, 347 (2002) 1860-1866. [11] P.L. Lin, J.L. Flynn, Understanding latent tuberculosis: a moving target, J Immunol, 185 (2010) 15-22. [12] American Thoracic Society, Diagnostic Standards and Classification of Tuberculosis in Adults and Children, American Journal of Respiratory and Critical Care Medicine, 161 (2000) 1376-1395. [13] P.-J. Cardona, A spotlight on liquefaction: evidence from clinical settings and experimental models in tuberculosis, Clin Dev Immunol, 2011 (2011). [14] N.A. Knechel, Tuberculosis: pathophysiology, clinical features, and diagnosis, Crit Care Nurse, 29 (2009) 34-43. [15] S. Luca, T. Mihaescu, History of BCG vaccine, Mædica, 8 (2013) 53-58. [16] P.W. Roche, J.A. Triccas, N. Winter, BCG vaccination against tuberculosis: past disappointments and future hopes, Trends Microbiol, 3 (1995) 397-401. [17] J.M. Pitt, S. Blankley, H. McShane, A. O'Garra, Vaccination against tuberculosis: how can we better BCG?, Microb Pathog, 58 (2013) 2-16. [18] J.A.C. Sterne, L.C. Rodrigues, I.N. Guedes, Does the efficacy of BCG decline with time since vaccination?, Int J Tuberc Lung Dis, 2 (1998) 200-207.

[19] I. Abubakar, L. Pimpin, C. Ariti, R. Beynon, P. Mangtani, J. Sterne, P. Fine, P. Smith, M. Lipman, D. Elliman, J. Watson, L. Drumright, P. Whiting, E. Vynnycky, L.C. Rodrigues, Systematic review and meta-analysis of the current evidence on the duration of protection by bacillus Calmette-Guérin vaccination against tuberculosis, Health Technol Assess, 17 (2013) 1-372. [20] M. Haile, G. Källenius, Recent developments in tuberculosis vaccines, Curr Opin Infect Dis, 18 (2005) 211-215. [21] S.N.M. Hanif, L. Garcia-Contreras, Pharmaceutical aerosols for the treatment and prevention of tuberculosis, Front Cell Infect Microbiol, 2 (2012) 118. [22] World Health Organization, Treatment of tuberculosis: guidelines. 4th edition, 2010. [23] C. Lienhardt, A. Vernon, M.C. Raviglione, New drugs and new regimens for the treatment of tuberculosis: review of the drug development pipeline and implications for national programmes, Curr Opin Pulm Med, 16 (2010) 186-193. [24] A. Zumla, P. Nahid, S.T. Cole, Advances in the development of new tuberculosis drugs and treatment regimens, Nat Rev Drug Discov, 12 (2013) 388-404. [25] A. Zumla, M. Raviglione, R. Hafner, C. Fordham von Reyn, Tuberculosis, N Engl J Med, 368 (2013) 745-755. [26] Das S, Tucker I, Stewart P, Inhaled dry powder formulations for treating tuberculosis, Curr Drug Deliv, 12 (2015) 26-39. [27] Garcia-Contreras L, Awashthi S, Hanif S. N. M., Hickey A. J., Inhaled vaccines for the prevention of tuberculosis, J Mycobac Dis, S1 (2012) 1-13. [28] A. Misra, A.J. Hickey, C. Rossi, G. Borchard, H. Terada, K. Makino, P.B. Fourie, P. Colombo, Inhaled drug therapy for treatment of tuberculosis, Tuberculosis, 91 (2011) 71-81. [29] A.S. Dharmadhikari, M. Kabadi, B. Gerety, A.J. Hickey, P.B. Fourie, E. Nardell, Phase I, single-dose, dose-escalating study of inhaled dry powder capreomycin: a new approach to therapy of drug-resistant tuberculosis, Antimicrob Agents Chemother, 57 (2013) 2613-2319. [30] J.G.Y. Chan, A.S. Tyne, A. Pang, A.J. McLachlan, V. Perera, J.C.Y. Chan, W.J. Britton, H.K. Chan, C.C. Duke, P.M. Young, D. Traini, Murine pharmacokinetics of rifapentine delivered as an inhalable dry powder, Int J Antimicrob Agents, 43 (2014) 319-323. [31] T. Hasegawa, K. Hirota, K. Tomoda, F. Ito, H. Inagawa, C. Kochi, G.-I. Soma, K. Makino, H. Terada, Phagocytic activity of alveolar macrophages toward polystyrene latex microspheres and PLGA microspheres loaded with anti-tuberculosis agent, Colloids Surf B Biointerfaces, 60 (2007) 221-228. [32] K. Hirota, T. Hasegawa, T. Nakajima, H. Inagawa, C. Kohchi, G.-I. Soma, K. Makino, H. Terada, Delivery of rifampicin–PLGA microspheres into alveolar macrophages is promising for treatment of tuberculosis, J Control Release, 142 (2010) 339-346. [33] K. Hirota, T. Hasegawa, T. Nakajima, K. Makino, H. Terada, Phagostimulatory effect of uptake of PLGA microspheres loaded with rifampicin on alveolar macrophages, Colloids Surf B Biointerfaces, 87 (2011) 293-298. [34] A. Gupta, A. Misra, V. Deretic, Targeted pulmonary delivery of inducers of host macrophage autophagy as a potential host-directed chemotherapy of tuberculosis, Adv Drug Deliv Rev, In Press (2016).

[35] A.B. Yadav, P. Muttil, A.K. Singh, R.K. Verma, M. Mohan, A.K. Agrawal, A.S. Verma, S.K. Sinha, A. Misra, Microparticles induce variable levels of activation in macrophages infected with Mycobacterium tuberculosis, Tuberculosis, 90 (2010) 188-196. [36] M. Bivas-Benita, T.H.M. Ottenhoff, H.E. Junginger, G. Borchard, Pulmonary DNA vaccination: concepts, possibilities and perspectives, J Control Release, 107 (2005) 1-29. [37] G. Diamond, D. Legarda, L.K. Ryan, The innate immune response of the respiratory epithelium, Immunol Rev, 173 (2000) 27-38. [38] P. Zhang, W.R. Summer, G.J. Bagby, S. Nelson, Innate immunity and pulmonary host defense, Immunol Rev, 173 (2000) 39-51. [39] A. Tyne, J.G.Y. Chan, E. Shanahan, I. Atmosukarto, H.-K. Chan, W. Britton, N. West, TLR2-targeted secreted proteins from Mycobacterium tuberculosis are protective as powdered pulmonary vaccines, Vaccine, 31 (2013) 4322-4329. [40] N.R. Labiris, M.B. Dolovich, Pulmonary drug delivery. Part I: physiological factors affecting therapeutic effectiveness of aerosolized medications, Br J Clin Pharmacol, 56 (2003) 588-599. [41] R.W. Dal Negro, Dry powder inhalers and the right things to remember: a concept review, Multidiscip Respir Med, 10 (2015) 13. [42] W.F. Tonnis, A.J. Lexmond, H.W. Frijlink, A.H. de Boer, W.L.J. Hinrichs, Devices and formulations for pulmonary vaccination, Expert Opin Drug Deliv, 10 (2013) 1383-1397. [43] J.L. Rau, The inhalation of drugs: advantages and problems, Respir Care, 50 (2005) 367- 382. [44] H.-K. Chan, Dry powder aerosol drug delivery-opportunities for colloid and surface scientists, Colloids Surf A, 284–285 (2006) 50-55. [45] W.W. Busse, T. Uryniak, Efficacy of inhaled budesonide (Pulmicort Turbuhaler®) in recent-onset moderate to severe persistent asthma, J Allergy Clin Immunol, 113 (2004) S119. [46] U.S. Food and Drug Administration, FDA approves Foradil Aerolizer (asthma treatment) (brief article), In Drug Topics, 145 (2001) 9. [47] M. Karhu, J. Kuikka, T. Kauppinen, K. Bergström, M. Vidgren, Pulmonary deposition of lactose carriers used in inhalation powders, Int J Pharm, 196 (2000) 95-103. [48] A.M. Healy, M.I. Amaro, K.J. Paluch, L. Tajber, Dry powders for oral inhalation free of lactose carrier particles, Adv Drug Deliv Rev, 75 (2014) 32-52. [49] S.P. Shah, A. Misra, Liposomal amikacin dry powder inhaler: effect of fines on in vitro performance, AAPS Pharm Sci Tech, 5 (2004) 107-113. [50] B.K. Padhi, M. B. Chougule, A. Misra, Aerosol performance of large respirable particles of amikacin sulfate produced by spray and freeze drying techniques, Curr Drug Deliv, 6 (2009) 8-16. [51] J. Fiegel, L. Garcia-Contreras, M. Thomas, J. VerBerkmoes, K. Elbert, A. Hickey, D. Edwards, Preparation and in vivo evaluation of a dry powder for inhalation of capreomycin, Pharm Res, 25 (2008) 805-811. [52] L. Garcia-Contreras, J. Fiegel, M.J. Telko, K. Elbert, A. Hawi, M. Thomas, J. VerBerkmoes, W.A. Germishuizen, P.B. Fourie, A.J. Hickey, D. Edwards, Inhaled large

70 porous particles of capreomycin for treatment of tuberculosis in a guinea pig model, Antimicrob Agents Chemother, 51 (2007) 2830-2836. [53] J.R. Manion, S.P. Cape, D.H. McAdams, L.G. Rebits, S. Evans, R.E. Sievers, Inhalable antibiotics manufactured through use of near-critical or supercritical fluids, Aerosol Sci Tech, 46 (2011) 403-410. [54] A. Schoubben, S. Giovagnoli, M.C. Tiralti, P. Blasi, M. Ricci, Capreomycin inhalable powders prepared with an innovative spray-drying technique, Int J Pharm, 469 (2014) 132-139. [55] C. Lawlor, M.P. O’Sullivan, B. Rice, P. Dillon, P.J. Gallagher, S. O’Leary, S. Shoyele, J. Keane, S.A. Cryan, Therapeutic aerosol bioengineering of targeted, inhalable microparticle formulations to treat Mycobacterium tuberculosis (MTb), J Mater Sci Mater Med, 23 (2012) 89-98. [56] S. Giovagnoli, P. Blasi, A. Schoubben, C. Rossi, M. Ricci, Preparation of large porous biodegradable microspheres by using a simple double-emulsion method for capreomycin sulfate pulmonary delivery, Int J Pharm, 333 (2007) 103-111. [57] A. Schoubben, P. Blasi, M.L. Marenzoni, L. Barberini, S. Giovagnoli, C. Cirotto, M. Ricci, Capreomycin supergenerics for pulmonary tuberculosis treatment: preparation, in vitro, and in vivo characterization, Eur J Pharm Biopharm, 83 (2013) 388-395. [58] R.K. Verma, W.A. Germishuizen, M.P. Motheo, A.K. Agrawal, A.K. Singh, M. Mohan, P. Gupta, U.D. Gupta, M. Cholo, R. Anderson, P.B. Fourie, A. Misra, Inhaled microparticles containing clofazimine are efficacious in treatment of experimental tuberculosis in mice, Antimicrob Agents Chemother, 57 (2013) 1050-1052. [59] W. Rojanarat, N. Changsan, E. Tawithong, S. Pinsuwan, H.-K. Chan, T. Srichana, Isoniazid proliposome powders for inhalation-preparation, characterization and cell culture studies, Int J Mol Sci, 12 (2011) 4414-4434. [60] A.J. Kundawala, H.A. Patel, V.A. Patel, D. Choudhary, Isoniazid loaded chitosan microspheres for pulmonary delivery: preparation and characterization, Der Pharmacia Sinica, 2 (2011) 88-97. [61] P.S. Pourshahab, K. Gilani, E. Moazeni, H. Eslahi, M.R. Fazeli, H. Jamalifar, Preparation and characterization of spray dried inhalable powders containing chitosan nanoparticles for pulmonary delivery of isoniazid, J Microencapsul, 28 (2011) 605-613. [62] W. Rojanarat, T. Nakpheng, E. Thawithong, N. Yanyium, T. Srichana, Levofloxacin- proliposomes: opportunities for use in lung tuberculosis, Pharmaceutics, 4 (2012) 385-412. [63] J.C. Sung, L. Garcia-Contreras, J.L. VerBerkmoes, C.A. Peloquin, K.J. Elbert, A.J. Hickey, D.A. Edwards, Dry powder nitroimidazopyran antibiotic PA-824 aerosol for inhalation, Antimicrob Agents Chemother, 53 (2009) 1338-1343. [64] L. Garcia-Contreras, J.C. Sung, P. Muttil, D. Padilla, M. Telko, J.L. VerBerkmoes, K.J. Elbert, A.J. Hickey, D.A. Edwards, Dry powder PA-824 aerosols for treatment of tuberculosis in guinea pigs, Antimicrob Agents Chemother, 54 (2010) 1436-1442. [65] S. Gad, L. Tajber, O.I. Corrigan, A.M. Healy, Preparation and characterisation of novel spray-dried nano-structured para-aminosalicylic acid particulates for pulmonary delivery: impact of ammonium carbonate on morphology, chemical composition and solid state, J Pharm Pharmacol, 64 (2012) 1264-1274.

[66] N. Tsapis, D. Bennett, K. O’Driscoll, K. Shea, M.M. Lipp, K. Fu, R.W. Clarke, D. Deaver, D. Yamins, J. Wright, C.A. Peloquin, D.A. Weitz, D.A. Edwards, Direct lung delivery of para- aminosalicylic acid by aerosol particles, Tuberculosis, 83 (2003) 379-385. [67] W. Rojanarat, T. Nakpheng, E. Thawithong, N. Yanyium, T. Srichana, Inhaled pyrazinamide proliposome for targeting alveolar macrophages, Drug Deliv, 19 (2012) 334-345. [68] C.A. Ober, L. Kalombo, H. Swai, R.B. Gupta, Preparation of rifampicin/lactose microparticle composites by a supercritical antisolvent-drug excipient mixing technique for inhalation delivery, Powder Technol, 236 (2013) 132-138. [69] Y.-J. Son, J.T. McConville, A new respirable form of rifampicin, Eur J Pharm Biopharm, 78 (2011) 366-376. [70] Y.-J. Son, J.T. McConville, Preparation of sustained release rifampicin microparticles for inhalation, J Pharm Pharmacol, 64 (2012) 1291-1302. [71] A. Kundawala, V. Patel, H. Patel, D. Choudhary, Preparation, in vitro characterization, and in vivo pharmacokinetic evaluation of respirable porous microparticles containing rifampicin, Sci Pharm, 82 (2014) 665-681. [72] A. J. Kundawala, V. A. Patel, H. V. Patel, D. Choudhary, Preparation of microparticles containing rifampicin as dry powder formulation: in vitro studies on aerosol performance., Am J Pharmtech Res, 2 (2012) 470-483. [73] T.V.P. Doan, J.C. Olivier, Preparation of rifampicin-loaded PLGA microspheres for lung delivery as aerosol by premix membrane homogenization, Int J Pharm, 382 (2009) 61-66. [74] T.V.P. Doan, W. Couet, J.C. Olivier, Formulation and in vitro characterization of inhalable rifampicin-loaded PLGA microspheres for sustained lung delivery, Int J Pharm, 414 (2011) 112-117. [75] N. Changsan, H.-K. Chan, F. Separovic, T. Srichana, Physicochemical characterization and stability of rifampicin liposome dry powder formulations for inhalation, J Pharm Sci, 98 (2009) 628-639. [76] N. Changsan, A. Nilkaeo, P. Pungrassami, T. Srichana, Monitoring safety of liposomes containing rifampicin on respiratory cell lines and in vitro efficacy against Mycobacterium bovis in alveolar macrophages, J Drug Target, 17 (2009) 751-762. [77] T. Ozeki, S. Beppu, T. Mizoe, Y. Takashima, H. Yuasa, H. Okada, Preparation of polymeric submicron particle-containing microparticles using a 4-fluid nozzle spray drier, Pharm Res, 23 (2006) 177-183. [78] K. Ohashi, T. Kabasawa, T. Ozeki, H. Okada, One-step preparation of rifampicin/poly(lactic-co-glycolic acid) nanoparticle-containing mannitol microspheres using a four-fluid nozzle spray drier for inhalation therapy of tuberculosis, J Control Release, 135 (2009) 19-24. [79] R. Diab, J. Brillault, A. Bardy, A.V.L. Gontijo, J.C. Olivier, Formulation and in vitro characterization of inhalable polyvinyl alcohol-free rifampicin-loaded PLGA microspheres prepared with sucrose palmitate as stabilizer: efficiency for ex vivo alveolar macrophage targeting, Int J Pharm, 436 (2012) 833-839. [80] I. Coowanitwong, V. Arya, P. Kulvanich, G. Hochhaus, Slow release formulations of inhaled rifampin, AAPS J, 10 (2008) 342-348.

[81] P. O'Hara, A. Hickey, Respirable PLGA microspheres containing rifampicin for the treatment of tuberculosis: manufacture and characterization, Pharm Res, 17 (2000) 955-961. [82] S. Suarez, P. O'Hara, M. Kazantseva, C. Newcomer, R. Hopfer, D. McMurray, A. Hickey, Airways delivery of rifampicin microparticles for the treatment of tuberculosis, J Antimicrob Chemother, 48 (2001) 431-434. [83] V. Patomchaiviwat, O. Paeratakul, P. Kulvanich, Formation of inhalable rifampicin- poly(L-lactide) microparticles by supercritical anti-solvent process, AAPS Pharm Sci Tech, 9 (2008) 1119-1129. [84] M.V. Vadakkan, K. Annapoorna, K.C. Sivakumar, S. Mundayoor, G.S.V. Kumar, Dry powder cationic lipopolymeric nanomicelle inhalation for targeted delivery of antitubercular drug to alveolar macrophage, Int J Nanomedicine, 8 (2013) 2871-2885. [85] L. Garcia Contreras, J. Sung, M. Ibrahim, K. Elbert, D. Edwards, A. Hickey, Pharmacokinetics of inhaled rifampicin porous particles for tuberculosis treatment: insight into rifampicin absorption from the lungs of guinea pigs, Mol Pharm, 12 (2015) 2642-2650. [86] R.V. Pai, R.R. Jain, A.S. Bannalikar, M.D. Menon, Development and evaluation of chitosan microparticles based dry powder inhalation formulations of rifampicin and rifabutin, J Aerosol Med Pulm Drug Deliv, 29 (2015) 179-195. [87] C. Liu, C. Kong, G. Wu, J. Zhu, B. Javid, F. Qian, Uniform and amorphous rifampicin microspheres obtained by freezing induced LLPS during lyophilization, Int J Pharm, 495 (2015) 500-507. [88] J.G.Y. Chan, C.C. Duke, H.X. Ong, J.C.Y. Chan, A.S. Tyne, H.-K. Chan, W.J. Britton, P.M. Young, D. Traini, A novel inhalable form of rifapentine, J Pharm Sci, 103 (2014) 1411- 1421. [89] A. Patil-Gadhe, V. Pokharkar, Single step spray drying method to develop proliposomes for inhalation: a systematic study based on quality by design approach, Pulm Pharmacol Ther, 27 (2014) 197-207. [90] A.A. Patil-Gadhe, A.Y. Kyadarkunte, M. Pereira, G. Jejurikar, M.S. Patole, A. Risbud, V.B. Pokharkar, Rifapentine-proliposomes for inhalation: in vitro and in vivo Toxicity, Toxicol Int, 21 (2014) 275-282. [91] T. Parumasivam, S.S.Y. Leung, D.H. Quan, J.A. Triccas, W.J. Britton, H.-K. Chan, Rifapentine-loaded PLGA microparticles for tuberculosis inhaled therapy: preparation and in vitro aerosol characterization, Eur J Pharm Sci, 88 (2016) 1-11. [92] J.G.Y. Chan, H.-K. Chan, C.A. Prestidge, J.A. Denman, P.M. Young, D. Traini, A novel dry powder inhalable formulation incorporating three first-line anti-tubercular antibiotics, Eur J Pharm Biopharm, 83 (2013) 285-292. [93] A. Bhardwaj, L. Kumar, R.K. Narang, R.S.R. Murthy, Development and characterization of ligand-appended liposomes for multiple drug therapy for pulmonary tuberculosis, Artif Cells Nanomed Biotechnol, 41 (2012) 52-59. [94] P. Muttil, J. Kaur, K. Kumar, A.B. Yadav, R. Sharma, A. Misra, Inhalable microparticles containing large payload of anti-tuberculosis drugs, Eur J Pharm Sci, 32 (2007) 140-150. [95] R.K. Verma, J.K. Mukker, R.S.P. Singh, K. Kumar, P.R.P. Verma, A. Misra, Partial biodistribution and pharmacokinetics of isoniazid and rifabutin following pulmonary delivery of inhalable microparticles to rhesus macaques, Mol Pharm, 9 (2012) 1011-1016.

[96] J. Chan, A. Tyne, A. Pang, H.-K. Chan, P. Young, W. Britton, C. Duke, D. Traini, A rifapentine-containing inhaled triple antibiotic formulation for rapid treatment of tubercular infection, Pharm Res, 31 (2013) 1239-1253. [97] T. Parumasivam, J.G.Y. Chan, A. Pang, D.H. Quan, J.A. Triccas, W.J. Britton, H.K. Chan, In vitro evaluation of inhalable verapamil-rifapentine particles for tuberculosis therapy, Mol Pharm, 13 (2016) 979-989. [98] T.H. Jin, E. Tsao, J. Goudsmit, V. Dheenadhayalan, J. Sadoff, Stabilizing formulations for inhalable powders of an adenovirus 35-vectored tuberculosis (TB) vaccine (AERAS-402), Vaccine, 28 (2010) 4369-4375. [99] Y.-L. Wong, S. Sampson, W.A. Germishuizen, S. Goonesekera, G. Caponetti, J. Sadoff, B.R. Bloom, D. Edwards, Drying a tuberculosis vaccine without freezing, Proc Natl Acad Sci USA, 104 (2007) 2591-2595. [100] L. Garcia-Contreras, Y.-L. Wong, P. Muttil, D. Padilla, J. Sadoff, J. DeRousse, W.A. Germishuizen, S. Goonesekera, K. Elbert, B.R. Bloom, R. Miller, P.B. Fourie, A. Hickey, D. Edwards, Immunization by a bacterial aerosol, Proc Natl Acad Sci, 105 (2008) 4656-4660. [101] D. Lu, L. Garcia-Contreras, P. Muttil, D. Padilla, D. Xu, J. Liu, M. Braunstein, D. McMurray, A. Hickey, Pulmonary immunization using antigen 85-B polymeric microparticles to boost tuberculosis immunity, AAPS J, 12 (2010) 338-347. [102] D. Lu, L. Garcia-Contreras, D. Xu, S. Kurtz, J. Liu, M. Braunstein, D. McMurray, A. Hickey, Poly (lactide-co-glycolide) microspheres in respirable sizes enhance an in vitro T cell response to recombinant Mycobacterium tuberculosis antigen 85B, Pharm Res, 24 (2007) 1834-1843. [103] P. Kuehl, A. Cherrington, D. Dobry, D. Edgerton, D. Friesen, C. Hobbs, C. Leach, B. Murri, D. Neal, D. Lyon, D. Vodak, M. Reed, Biologic comparison of inhaled insulin formulations: ExuberaTM and novel spray-dried engineered particles of dextran-10, AAPS PharmSciTech, 15 (2014) 1545-1550. [104] D.R. VanDevanter, D.E. Geller, Tobramycin administered by the TOBI(®) Podhaler(®) for persons with cystic fibrosis: a review, Med Devices 4(2011) 179-188. [105] J.D. Brannan, S.D. Anderson, C.P. Perry, R. Freed-Martens, A.R. Lassig, B. Charlton, G. the Aridol Study, The safety and efficacy of inhaled dry powder mannitol as a bronchial provocation test for airway hyperresponsiveness: a phase 3 comparison study with hypertonic (4.5%) saline, Respir Res, 6 (2005) 144-144. [106] R. Vehring, Pharmaceutical particle engineering via spray drying, Pharm Res, 25 (2008) 999-1022. [107] Q. Zhou, S.S.Y. Leung, P. Tang, T. Parumasivam, Z.H. Loh, H.-K. Chan, Inhaled formulations and pulmonary drug delivery systems for respiratory infections, Adv Drug Deliv Rev, 85 (2015) 83-99. [108] N. Chew, B.Y. Shekunov, H.H. Tong, A.H. Chow, C. Savage, J. Wu, H.-K. Chan, Effect of amino acids on the dispersion of disodium cromoglycate powders, J Pharm Sci, 94 (2005) 2289-2300. [109] H.-Y. Li, H. Neill, R. Innocent, P. Seville, I. Williamson, J.C. Birchall, Enhanced dispersibility and deposition of spray-dried powders for pulmonary gene therapy, J Drug Target, 11 (2003) 425-432.

[110] J. Raula, A. Lähde, E.I. Kauppinen, Aerosolization behavior of carrier-free L-leucine coated salbutamol sulphate powders, Int J Pharm, 365 (2009) 18-25. [111] J. Raula, J. Seppälä, J. Malm, M. Karppinen, E. Kauppinen, Structure and dissolution of L-Leucine-coated salbutamol sulphate aerosol particles, AAPS PharmSciTech, 13 (2012) 707- 712. [112] Q. Zhou, L. Qu, T. Gengenbach, I. Larson, P.J. Stewart, D.A.V. Morton, Effect of surface coating with magnesium stearate via mechanical dry powder coating approach on the aerosol performance of micronized drug powders from dry powder inhalers, AAPS PharmSciTech, 14 (2013) 38-44. [113] C. Parlati, P. Colombo, F. Buttini, P. Young, H. Adi, A. Ammit, D. Traini, Pulmonary spray dried powders of tobramycin containing sodium stearate to improve aerosolization efficiency, Pharm Res, 26 (2009) 1084-1092. [114] D.V. Gowda, D.M. Vishnu, S. Meenakshi, H. Siddaramaiah, K.G.N. Vishal, Formulation and evaluation of dry powders containing anti tuberculosis drugs for pulmonary delivery, IAJPR, 3 (2013) 1239-1248. [115] L.L.I.J. Booysen, L. Kalombo, E. Brooks, R. Hansen, J. Gilliland, V. Gruppo, P. Lungenhofer, B. Semete-Makokotlela, H.S. Swai, A.F. Kotze, A. Lenaerts, L.H. du Plessis, In vivo/in vitro pharmacokinetic and pharmacodynamic study of spray-dried poly-(dl-lactic-co- glycolic) acid nanoparticles encapsulating rifampicin and isoniazid, Int J Pharm, 444 (2013) 10-17. [116] P.B. Fourie, W.A. Germishuizen, Y.L. Wong, D.A. Edwards, Spray drying TB vaccines for pulmonary administration, Expert Opin Biol Ther, 8 (2008) 857-863. [117] Elservier, Capreomycin, Tuberculosis 88 (2008) 89-91. [118] S. Shin, J. Furin, F. Alcantara, A. Hyson, K. Joseph, E. Sanchez, M. Rich, Hypokalemia among patients receiving treatment for multidrug-resistant tuberculosis, Anglais, 125 (2004) 974-980. [119] D. Heng, S.H. Lee, W.K. Ng, R.B.H. Tan, The nano spray dryer B-90, Expert Opin Drug Deliv, 8 (2011) 965-972. [120] J. van Ingen, R.E. Aarnoutse, P.R. Donald, A.H. Diacon, R. Dawson, G. Plemper van Balen, S.H. Gillespie, M.J. Boeree, Why do we use 600 mg of rifampicin in tuberculosis treatment?, Clin Infect Dis, 52 (2011) e194-e199. [121] T. Ozeki, T. Tagami, Drug/polymer nanoparticles prepared using unique spray nozzles and recent progress of inhaled formulation, Asian J Pharm Sci, 9 (2014) 236-243. [122] T. Mizoe, T. Ozeki, H. Okada, Application of a four-fluid nozzle spray drier to prepare inhalable rifampicin-containing mannitol microparticles, AAPS PharmSciTech, 9 (2008) 755- 761. [123] F.A. Sirgel, P.B. Fourie, P.R. Donald, N. Padayatchi, R. Rustomjee, J. Levin, G. Roscigno, J. Norman, H. McIlleron, D.A. Mitchison, The early bactericidal activities of rifampin and rifapentine in pulmonary tuberculosis, Am J Respir Crit Care Med, 172 (2005) 128-135. [124] ClinicalTrials.gov, Assessing PA-824 for tuberculosis (the APT trial), in, U.S. National Institutes of Health, https://clinicaltrials.gov/ct2/show/record/NCT02256696, 2016.

[125] H.I. Seifart, D.P. Parkin, P.R. Donald, Stability of isoniazid, rifampin and pyrazinamide in suspensions used for the treatment of tuberculosis in children, Pediatr Infect Dis J, 10 (1991) 827-831. [126] R. Sharma, D. Saxena, A. Dwivedi, A. Misra, Inhalable microparticles containing drug combinations to target alveolar macrophages for treatment of pulmonary tuberculosis, Pharm Res, 18 (2001) 1405-1410. [127] D. McAdams, D. Chen, D. Kristensen, Spray drying and vaccine stabilization, Expert Rev Vaccines, 11 (2012) 1211-1219. [128] V. Naini, P.R. Byron, E.M. Phillips, Physicochemical stability of crystalline sugars and their spray-dried forms: dependence upon relative humidity and suitability for use in powder inhalers, Drug Dev Ind Pharm, 24 (1998) 895-909. [129] B. Burrowes, D.R. Harper, J. Anderson, M. McConville, M.C. Enright, Bacteriophage therapy: potential uses in the control of antibiotic-resistant pathogens, Expert Rev Anti Infect Ther, 9 (2011) 775-785. [130] S.T. Abedon, C. Thomas-Abedon, Phage therapy pharmacology, Curr Pharm Biotechnol, 11 (2010) 28-47. [131] A.D. Morgan, M.B. Bonsall, A. Buckling, Impact of bacterial mutation rate on coevolutionary dynamics between bacteria and phages, Evolution, 64 (2010) 2980-2987. [132] Delacoste P., Considerations on the treatment of common respiratory diseases by means of bacteriophages, Rev Med Suisse Romande, 79 (1959) 552-563. [133] A. Sulakvelidze, Z. Alavidze, J.G. Morris, Bacteriophage therapy, Antimicrob Agents Chemother, 45 (2001) 649-659. [134] K.-Y. Liu, W.-H. Yang, X.-K. Dong, L.-M. Cong, N. Li, Y. Li, Z.-B. Wen, Z. Yin, Z.-J. Lan, W.-P. Li, J.-S. Li, Inhalation study of mycobacteriophage D29 aerosol for mice by endotracheal route and nose-only exposure, J Aerosol Med Pulm Drug Deliv, (2016). [135] L. Li, S. Sun, T. Parumasivam, J.A. Denman, T. Gengenbach, P. Tang, S. Mao, H.-K. Chan, l-Leucine as an excipient against moisture on in vitro aerosolization performances of highly hygroscopic spray-dried powders, Eur J Pharm Biopharm, 102 (2016) 132-141. [136] S. Matinkhoo, K.H. Lynch, J.J. Dennis, W.H. Finlay, R. Vehring, Spray-dried respirable powders containing bacteriophages for the treatment of pulmonary infections, J Pharm Sci, 100 (2011) 5197-5205. [137] D. Vandenheuvel, A. Singh, K. Vandersteegen, J. Klumpp, R. Lavigne, G. Van den Mooter, Feasibility of spray drying bacteriophages into respirable powders to combat pulmonary bacterial infections, Eur J Pharm Biopharm, 84 (2013) 578-582. [138] L. Golshahi, K.H. Lynch, J.J. Dennis, W.H. Finlay, In vitro lung delivery of bacteriophages KS4-M and ΦKZ using dry powder inhalers for treatment of Burkholderia cepacia complex and Pseudomonas aeruginosa infections in cystic fibrosis, J Appl Microbiol, 110 (2011) 106-117. [139] S.S.Y. Leung, T. Parumasivam, F.G. Gao, N.B. Carrigy, R. Vehring, W.H. Finlay, S. Morales, W.J. Britton, E. Kutter, H.-K. Chan, Production of inhalation phage powders using spray freeze drying and spray drying techniques for treatment of respiratory infections, Pharm Res, 33 (2016) 1486-1496.

[140] P. Knezevic, D. Obreht, S. Curcin, M. Petrusic, V. Aleksic, R. Kostanjsek, O. Petrovic, Phages of Pseudomonas aeruginosa: response to environmental factors and in vitro ability to inhibit bacterial growth and biofilm formation, J Appl Microbiol, 111 (2011) 245-254. [141] S. Hoe, D.D. Semler, A.D. Goudie, K.H. Lynch, S. Matinkhoo, W.H. Finlay, J.J. Dennis, R. Vehring, Respirable bacteriophages for the treatment of bacterial lung infections, J Aerosol Med Pulm Drug Deliv, 26 (2013) 317-335. [142] S. Branston, E. Stanley, J. Ward, E. Keshavarz-Moore, Study of robustness of filamentous bacteriophages for industrial applications, Biotechnol Bioeng, 108 (2011) 1468- 1472. [143] D. Vandenheuvel, J. Meeus, R. Lavigne, G. Van den Mooter, Instability of bacteriophages in spray-dried trehalose powders is caused by crystallization of the matrix, Int J Pharm, 472 (2014) 202-205. [144] A. Misra, K. Jinturkar, D. Patel, J. Lalani, M. Chougule, Recent advances in liposomal dry powder formulations: preparation and evaluation, Expert Opin Drug Deliv, 6 (2009) 71- 89. [145] M.E. Ali, A. Lamprecht, Spray freeze drying for dry powder inhalation of nanoparticles, Eur J Pharm Biopharm, 87 (2014) 510-517. [146] Z.L. Wang, W.H. Finlay, M.S. Peppler, L.G. Sweeney, Powder formation by atmospheric spray-freeze-drying, Powder Technol, 170 (2006) 45-52. [147] R.J. Malcolmson, J.K. Embleton, Dry powder formulations for pulmonary delivery, Pharm Sci Tech Today, 1 (1998) 394-398. [148] K.M. Sharif, M.M. Rahman, J. Azmir, A. Mohamed, M.H.A. Jahurul, F. Sahena, I.S.M. Zaidul, Experimental design of supercritical fluid extraction – a review, J Food Process Eng, 124 (2014) 105-116. [149] Y. Sun, Supercritical fluid particle design of DPI formulations (Review), Curr Pharm Des, 21 (2015) 2516-2542. [150] S.P. Cape, J.A. Villa, E.T.S. Huang, T.-H. Yang, J.F. Carpenter, R.E. Sievers, Preparation of active proteins, vaccines and pharmaceuticals as fine powders using supercritical or near-critical fluids, Pharm Res, 25 (2008) 1967-1990. [151] G. Sessa, G. Weissmann, Phospholipid spherules (liposomes) as a model for biological membranes, J Lipid Res, 9 (1968) 310-318. [152] A.D. Bangham, R.W. Horne, Negative staining of phospholipids and their structural modification by surface-active agents as observed in the electron microscope, J Mol Biol, 8 (1964) 660-668. [153] S. Giovagnoli, P. Blasi, C. Vescovi, G. Fardella, I. Chiappini, L. Perioli, M. Ricci, C. Rossi, Unilamellar vesicles as potential capreomycin sulfate carriers: preparation and physicochemical characterization, AAPS PharmSciTech, 4 (2003) 549-560. [154] Samad A, Sultana Y, Aqil M, Liposomal drug delivery systems: an update review, Curr Drug Deliv, 4 (2007) 297-305. [155] D. Cipolla, I. Gonda, H.-K. Chan, Liposomal formulations for inhalation, Ther Deliv, 4 (2013) 1047-1072. [156] D. Cipolla, J. Blanchard, I. Gonda, Development of liposomal ciprofloxacin to treat lung infections, Pharmaceutics, 8 (2016) 6. 77

[157] M.M. Gaspar, U. Bakowsky, C. Ehrhardt, Inhaled liposomes - current strategies and future challenges, J Biomed Nanotechnol, 4 (2008) 245-257. [158] A. Gürsoy, E. Kut, S. Özkırımlı, Co-encapsulation of isoniazid and rifampicin in liposomes and characterization of liposomes by derivative spectroscopy, Int J Pharm, 271 (2004) 115-123. [159] N.K. Jain, V. Mishra, N.K. Mehra, Targeted drug delivery to macrophages, Expert Opin Drug Deliv, 10 (2013) 353-367. [160] A. Agarwal, H. Kandpal, H.P. Gupta, N.B. Singh, C.M. Gupta, Tuftsin-bearing liposomes as rifampin vehicles in treatment of tuberculosis in mice, Antimicrob Agents Chemother, 38 (1994) 588-593. [161] M. H. Al-Hallak, M. K. Sarfraz, S. Azarmi, W. H. Roa, W. H. Finlay, R. Löbenberg, Pulmonary delivery of inhalable nanoparticles: dry powder inhalers, Ther Deliv, 2 (2011) 1313-1324. [162] S.A. Agnihotri, K.S. Soppimath, G.V. Betageri, Controlled release application of multilamellar vesicles: a novel drug delivery approach, Drug Deliv, 17 (2010) 92-101. [163] U.S. Pharmacopeia, Powder flow, in: P. Forum (Ed.) http://www.pharmacopeia.cn/v29240/usp29nf24s0_c1174.html, 2008-2010. [164] W.B. Liechty, D.R. Kryscio, B.V. Slaughter, N.A. Peppas, Polymers for drug delivery systems, Annu Rev Chem Biomol Eng, 1 (2010) 149-173. [165] J.U. Menon, P. Ravikumar, A. Pise, D. Gyawali, C.C.W. Hsia, K.T. Nguyen, Polymeric nanoparticles for pulmonary protein and DNA delivery, Acta Biomater, 10 (2014) 2643-2652. [166] H.K. Makadia, S.J. Siegel, Poly lactic-co-glycolic acid (PLGA) as biodegradable controlled drug delivery carrier, Polymers (Basel), 3 (2011) 1377-1397. [167] D.L. Clemens, B.-Y. Lee, M. Xue, C.R. Thomas, H. Meng, D. Ferris, A.E. Nel, J.I. Zink, M.A. Horwitz, Targeted intracellular delivery of antituberculosis drugs to Mycobacterium tuberculosis-infected macrophages via functionalized mesoporous silica nanoparticles, Antimicrob Agents Chemother, 56 (2012) 2535-2545. [168] C. Gottstein, G. Wu, B.J. Wong, J.A. Zasadzinski, Precise quantification of nanoparticle internalization, ACS Nano, 7 (2013) 4933-4945. [169] S.M. Abdelghany, D. Schmid, J. Deacon, J. Jaworski, F. Fay, K.M. McLaughlin, J.A. Gormley, J.F. Burrows, D.B. Longley, R.F. Donnelly, C.J. Scott, Enhanced antitumor activity of the photosensitizer meso-Tetra(N-methyl-4-pyridyl) porphine tetra tosylate through encapsulation in antibody-targeted chitosan/alginate nanoparticles, Biomacromolecules, 14 (2013) 302-310. [170] S.M. Abdelghany, D.J. Quinn, R.J. Ingram, B.F. Gilmore, R.F. Donnelly, C.C. Taggart, C.J. Scott, Gentamicin-loaded nanoparticles show improved antimicrobial effects towards Pseudomonas aeruginosa infection, Int J Nanomedicine, 7 (2012) 4053-4063. [171] J. Deacon, S.M. Abdelghany, D.J. Quinn, D. Schmid, J. Megaw, R.F. Donnelly, D.S. Jones, A. Kissenpfennig, J.S. Elborn, B.F. Gilmore, C.C. Taggart, C.J. Scott, Antimicrobial efficacy of tobramycin polymeric nanoparticles for Pseudomonas aeruginosa infections in cystic fibrosis: formulation, characterisation and functionalisation with dornase alfa (DNase), J Control Release, 198 (2015) 55-61.

[172] R.F. Donnelly, D.I.J. Morrow, F. Fay, C.J. Scott, S. Abdelghany, R.R.T. Singh, M.J. Garland, A. David Woolfson, Microneedle-mediated intradermal nanoparticle delivery: potential for enhanced local administration of hydrophobic pre-formed photosensitisers, Photodiagnosis Photodyn Ther, 7 (2010) 222-231. [173] D. Wei, R. Dave, R. Pfeffer, Mixing and characterization of nanosized powders: an assessment of different techniques, J Nanopart Res, 4 (2002) 21-41. [174] M. Knoch, M. Keller, The customised electronic nebuliser: a new category of liquid aerosol drug delivery systems, Expert Opin Drug Deliv, 2 (2005) 377-390. [175] A. Ari, Jet, ultrasonic, and mesh nebulizers: an evaluation of nebulizers for better clinical outcomes, Eurasian J Pulmonol, 16 (2014) 1-7. [176] Q. Zhou, P. Tang, S.S.Y. Leung, J.G.Y. Chan, H.-K. Chan, Emerging inhalation aerosol devices and strategies: where are we headed?, Adv Drug Deliv Rev, 75 (2014) 3-17. [177] M. Hoppentocht, P. Hagedoorn, H.W. Frijlink, A.H. de Boer, Developments and strategies for inhaled antibiotic drugs in tuberculosis therapy: a critical evaluation, Eur J Pharm Biopharm, 86 (2014) 23-30. [178] M.R. Elkins, P. Robinson, S.D. Anderson, C.P. Perry, E. Daviskas, B. Charlton, Inspiratory flows and volumes in subjects with cystic fibrosis using a new dry powder inhaler device, Open Respir Med J, 8 (2014) 1-7. [179] R. Hodder, D. Price, Patient preferences for inhaler devices in chronic obstructive pulmonary disease: experience with Respimat(®) Soft Mist(™) Inhaler, Int J Chron Obstruct Pulmon Dis, 4 (2009) 381-390. [180] A.C. Moore, S. Stone, Meeting the needs of patients with COPD: patients' preference for the Diskus inhaler compared with the Handihaler, Int J Clin Pract, 58 (2004) 444-450. [181] Q. Zhou, Z. Tong, P. Tang, M. Citterio, R. Yang, H.-K. Chan, Effect of device design on the aerosolization of a carrier-based dry powder inhaler-a case study on Aerolizer(®) Foradile (®), AAPS J, 15 (2013) 511-522. [182] S.S.Y. Leung, T. Parumasivam, C. Mauro, H.-K. Chan, A modified Aerolizer® for delivery of high dose dry powder therapeutics, in: Respiratory Drug Delivery (RDD®) 2014 Fajardo, Puerto Rico, 2014. [183] P.M. Young, J. Crapper, G. Philips, K. Sharma, H.-K. Chan, D. Traini, Overcoming dose limitations using the Orbital® multi-breath dry powder inhaler, J Aerosol Med Pulm Drug Deliv, 27 (2013) 138-147. [184] P.M. Young, R.O. Salama, B. Zhu, G. Phillips, J. Crapper, H.-K. Chan, D. Traini, Multi- breath dry powder inhaler for delivery of cohesive powders in the treatment of bronchiectasis, Drug Dev Ind Pharm, 41 (2014) 859-865. [185] PA Knowledge Limited, The Orbital® inhaler, developed by PA for Pharmaxis, is being used in key research to combat tuberculosis, in: http://www.paconsulting.com/our- experience/an-inhaler-orbital-that-pa-developed-for-pharmaxis-is-being-used-in-key- research-to-combat-tuberculosis/ (Ed.), 2015. [186] J.G.Y. Chan, J. Wong, Q.T. Zhou, S.S.Y. Leung, H.-K. Chan, Advances in device and formulation technologies for pulmonary drug delivery, AAPS Pharm Sci Tech, 15 (2014) 882- 897.

[187] A.H. de Boer, P. Hagedoorn, E.M. Westerman, P.P.H. Le Brun, H.G.M. Heijerman, H.W. Frijlink, Design and in vitro performance testing of multiple air classifier technology in a new disposable inhaler concept (Twincer®) for high powder doses, Eur J Pharm Sci, 28 (2006) 171- 178. [188] M. Hoppentocht, O.W. Akkerman, P. Hagedoorn, H.W. Frijlink, A.H. de Boer, The Cyclops for pulmonary delivery of aminoglycosides; a new member of the Twincer™ family, Eur J Pharm Biopharm, 90 (2015) 8-15. [189] E.A. Bronsky, J. Grossman, M.J. Henis, P.P. Gallo, Ü. Yegen, G.D. Cioppa, J. Kottakis, S. Mehra, Inspiratory flow rates and volumes with the Aerolizer dry powder inhaler in asthmatic children and adults, Curr Med Res Opin, 20 (2004) 131-137. [190] L.P. Malmberg, P. Rytilä, P. Happonen, T. Haahtela, Inspiratory flows through dry powder inhaler in chronic obstructive pulmonary disease: age and gender rather than severity matters, Int J Chron Obstruct Pulmon Dis, 5 (2010) 257-262. [191] N. Islam, E. Gladki, Dry powder inhalers (DPIs) - a review of device reliability and innovation, Int J Pharm, 360 (2008) 1-11. [192] S.P. Newman, Dry powder inhalers for optimal drug delivery, Expert Opin Biol Ther, 4 (2004) 23-33. [193] T.E. Corcoran, R. Venkataramanan, R.M. Hoffman, M.P. George, A. Petrov, T. Richards, S. Zhang, J. Choi, Y.Y. Gao, C.D. Oakum, R.O. Cook, M. Donahoe, Systemic delivery of atropine sulfate by the MicroDose Dry-Powder Inhaler, J Aerosol Med Pulm Drug Deliv, 26 (2013) 46-55. [194] M. Ibrahim, R. Verma, L. Garcia-Contreras, Inhalation drug delivery devices: technology update, Med Devices (Auckl), 8 (2015) 131-139. [195] G. Hagan, N. Nathani, Clinical review: tuberculosis on the intensive care unit, Crit Care, 17 (2013) 240. [196] P. Tang, H.-K. Chan, D. Rajbhandari, P. Phipps, Method to introduce mannitol powder to intubated patients to improve sputum clearance, J Aerosol Med Pulm Drug Deliv, 24 (2011) 1-9. [197] S.S.Y. Leung, T. Parumasivam, P. Tang, H.-K. Chan, A proof-of-principle setup for delivery of Relenza® (Zanamivir) inhalation powder to intubated patients, J Aerosol Med Pulm Drug Deliv, 29 (2015) 30-35. [198] H.-K. Chan, D. Rajbhandari, P. Tang, J.D. Brannan, P. Phipps, Safety of administering dry powder mannitol to stimulate sputum clearance in intubated intensive care patients with sputum retention: a pilot study, in: B105. Devices and monitoring in the intensive care unit, American Thoracic Society, 2012, pp. A6809-A6809. [199] C.J. Riendeau, H. Kornfeld, THP-1 cell apoptosis in response to mycobacterial infection, Infect Immun, 71 (2003) 254-259. [200] J.S. Patil, V.K. Devi, K. Devi, S. Sarasija, A novel approach for lung delivery of rifampicin-loaded liposomes in dry powder form for the treatment of tuberculosis, Lung India, 32 (2015) 331-338. [201] M. Sakagami, In vivo, in vitro and ex vivo models to assess pulmonary absorption and disposition of inhaled therapeutics for systemic delivery, Adv Drug Deliv Rev, 58 (2006) 1030- 1060.

[202] A. Steimer, E. Haltner, C.M. Lehr, Cell culture models of the respiratory tract relevant to pulmonary drug delivery, J Aerosol Med, 18 (2005) 137-182. [203] J.J. Salomon, P. Galeron, N. Schulte, P.R. Morow, D. Severynse-Stevens, H. Huwer, N. Daum, C.-M. Lehr, A.J. Hickey, C. Ehrhardt, Biopharmaceutical in vitro characterization of CPZEN-45, a drug candidate for inhalation therapy of tuberculosis, Ther Deliv, 4 (2013) 915- 923. [204] D.N. McMurray, F.M. Collins, A.M. Dannenberg, Jr., D.W. Smith, Pathogenesis of experimental tuberculosis in animal models, in: T. Shinnick (Ed.) Tuberculosis, Springer Berlin Heidelberg, 1996, pp. 157-179. [205] L. Garcia-Contreras, V. Sethuraman, M. Kazantseva, V. Godfrey, A.J. Hickey, Evaluation of dosing regimen of respirable rifampicin biodegradable microspheres in the treatment of tuberculosis in the guinea pig, J Antimicrob Chemother, 58 (2006) 980-986. [206] R. Parikh, L. Patel, S. Dalwadi, Microparticles of rifampicin: comparison of pulmonary route with oral route for drug uptake by alveolar macrophages, phagocytosis activity and toxicity study in albino rats, Drug Delivery, 21 (2013) 406-411. [207] D.N. McMurray, A nonhuman primate model for preclinical testing of new tuberculosis vaccines, Clin Infect Dis, 30 (2000) S210-S212. [208] H.R. Black., R.S. Griffith, A.M. Peabody, Absorption, excretion and metabolism of capreomycin in normal and diseased states, Ann N Y Acad Sci, 135 (1966) 974-982. [209] D.E. Geller, M.W. Konstan, J. Smith, S.B. Noonberg, C. Conrad, Novel tobramycin inhalation powder in cystic fibrosis subjects: Pharmacokinetics and safety, Pediatr Pulmonol, 42 (2007) 307-313. [210] M.T. Newhouse, P.H. Hirst, S.P. Duddu, Y.H. Walter, T.E. Tarara, A.R. Clark, J.G. Weers, Inhalation of a dry powder tobramycin PulmoSphere formulation in healthy volunteers, Anglais, 124 (2003) 360-366. [211] M. Hoppentocht, O.W. Akkerman, P. Hagedoorn, J.-W.C. Alffenaar, T.S. van der Werf, H.A.M. Kerstjens, H.W. Frijlink, A.H. de Boer, Tolerability and pharmacokinetic evaluation of inhaled dry powder tobramycin free base in non-cystic fibrosis bronchiectasis patients, PLoS ONE, 11 (2016) e0149768. [212] ClinicalTrials.gov, The safety of tuberculosis treatments by oral inhalation, in, U.S. National Institutes of Health, https://clinicaltrials.gov/ct2/show/record/NCT01785927, 2016. [213] S. Subbian, L. Tsenova, M.-J. Kim, H.C. Wainwright, A. Visser, N. Bandyopadhyay, J.S. Bader, P.C. Karakousis, G.B. Murrmann, L.-G. Bekker, D.G. Russell, G. Kaplan, Lesion- specific immune response in granulomas of patients with pulmonary tuberculosis: a pilot study, PLoS ONE, 10 (2015) e0132249. [214] S.J. Sasindran, J.B. Torrelles, Mycobacterium tuberculosis infection and inflammation: what is beneficial for the host and for the bacterium?, Front Microbiol, 2 (2011) 2. [215] R. Sharma, P. Muttil, A.B. Yadav, S.K. Rath, V.K. Bajpai, U. Mani, A. Misra, Uptake of inhalable microparticles affects defence responses of macrophages infected with Mycobacterium tuberculosis H37Ra, J Antimicrob Chemother, 59 (2007) 499-506. [216] A.B. Yadav, A. Misra, Enhancement of apoptosis of THP-1 cells infected with Mycobacterium tuberculosis by inhalable microparticles and relevance to bactericidal activity, Antimicrob Agents Chemother, 51 (2007) 3740-3742.

[217] A.B. Yadav, R. Sharma, P. Muttil, A.K. Singh, R.K. Verma, M. Mohan, S.K. Patel, A. Misra, Inhalable microparticles containing isoniazid and rifabutin target macrophages and 'stimulate the phagocyte' to achieve high efficacy, Indian J Exp Biol, 47 (2009) 469-474. [218] A.B. Yadav, A.K. Singh, R.K. Verma, M. Mohan, A.K. Agrawal, A. Misra, The devil’s advocacy: when and why inhaled therapies for tuberculosis may not work, Tuberculosis, 91 (2011) 65-66. [219] L.M. Hodge, M. Marinaro, H.P. Jones, J.R. McGhee, H. Kiyono, J.W. Simecka, Immunoglobulin A (IgA) responses and IgE-associated inflammation along the respiratory tract after mucosal but not systemic immunization, Infect Immun, 69 (2001) 2328-2338. [220] J.W. Simecka, R.J. Jackson, H. Kiyono, J.R. McGhee, Mucosally induced immunoglobulin E-associated inflammation in the respiratory tract, Infect Immun, 68 (2000) 672-679. [221] D.A. Hokey, A. Misra, Aerosol vaccines for tuberculosis: a fine line between protection and pathology, Tuberculosis, 91 (2011) 82-85. [222] D. Kobasa, S.M. Jones, K. Shinya, J.C. Kash, J. Copps, H. Ebihara, Y. Hatta, J. Hyun Kim, P. Halfmann, M. Hatta, F. Feldmann, J.B. Alimonti, L. Fernando, Y. Li, M.G. Katze, H. Feldmann, Y. Kawaoka, Aberrant innate immune response in lethal infection of macaques with the 1918 influenza virus, Nature, 445 (2007) 319-323. [223] J. H. Anthony, A. Misra, P.B. Fourie, Dry powder antibiotic aerosol product development: inhaled therapy for tuberculosis, J Pharm Sci, 102 (2013) 3900-3907. [224] C. Gervelas, A.L. Serandour, S. Geiger, G. Grillon, P. Fritsch, C. Taulelle, B. Le Gall, H. Benech, J.R. Deverre, E. Fattal, N. Tsapis, Direct lung delivery of a dry powder formulation of DTPA with improved aerosolization properties: effect on lung and systemic decorporation of plutonium, J Control Release, 118 (2007) 78-86. [225] M. Bivas-Benita, M. Oudshoorn, S. Romeijn, K. van Meijgaarden, H. Koerten, H. van der Meulen, G. Lambert, T. Ottenhoff, S. Benita, H. Junginger, G. Borchard, Cationic submicron emulsions for pulmonary DNA immunization, J Control Release, 100 (2004) 145- 155. [226] S. Prior, B. Gander, N. Blarer, H.P. Merkle, M.a.L. Subirá, J.M. Irache, C. Gamazo, In vitro phagocytosis and monocyte-macrophage activation with poly(lactide) and poly(lactide- co-glycolide) microspheres, Eur J Pharm Sci, 15 (2002) 197-207. [227] J.G.Y. Chan, Inhaled rifamycin dry powders targeting tuberculosis (thesis), in: School of Medical Sciences, University of Sydney, 2014. [228] J.G.Y. Chan, X. Bai, D. Traini, An update on the use of rifapentine for tuberculosis therapy, Expert Opin Drug Deliv, 11 (2014) 421-431.

CHAPTER 2

IN VITRO EVALUATION OF INHALABLE

VERAPAMIL-RIFAPENTINE PARTICLES FOR

TUBERCULOSIS THERAPY

This chapter has been published in Molecular Pharmaceutics, 2016; 13: 979-989, with the title

‘In vitro evaluation of inhalable verapamil-rifapentine particles for tuberculosis therapy’.

Authors: Parumasivam T, Chan JGY, Pang A, Quan DH, Triccas JA, Britton WJ, Chan H-K.

ABSTRACT

Recent studies have demonstrated that drug efflux pumps of M. tuberculosis provide a crucial mechanism in the development of drug resistant to anti-mycobacterial drugs. Drugs that inhibit these efflux pumps, such as verapamil have shown the potential in enhancing the treatment success. We therefore hypothesised that the combined inhaled administration of verapamil and a first-line rifamycin antibiotic will further improve the treatment efficacy. An inhalable dry powder consisting of amorphous verapamil and crystalline rifapentine with L-leucine as an excipient was produced by spray drying. The in vitro aerosol characteristic of the powder, its microbiological activity and stability were assessed. When the powder was dispersed by an

® Osmohaler , the total fine particle fraction (FPFtotal, wt. % of particles in aerosol < 5 µm) of verapamil and rifapentine was 77.4 ± 1.1 % and 71.5 ± 2.0 %, respectively. The combination drug formulation showed a minimum inhibitory concentration (MIC90) similar to that of rifapentine alone when tested against both M. tuberculosis H37Ra and M. tuberculosis H37Rv strains. Importantly, the combination resulted in increased killing of M. tuberculosis H37Ra within the infected macrophage cells compared to either verapamil or rifapentine alone. In assessing cellular toxicity, the combination exhibited an acceptable half maximal inhibitory concentration (IC50) values (62.5 µg/mL) on both human monocytic (THP-1) and lung alveolar basal epithelial (A549) cell lines. Finally, the powder was stable after 3 months storage in 0 % relative humidity at 20 ± 3°C.

1. INTRODUCTION

Tuberculosis (TB) caused by the intracellular pathogen Mycobacterium tuberculosis is an ancient lung infectious disease that remains a major cause of global mortality. In 2013, the

World Health Organization (WHO) reported that there were 1.5 million deaths attributed to the disease and 6.1 million people with active infection [1] with a further estimated 2 billion people have latent TB [2]. In addition, there are alarming rates of multi-drug resistant (MDR) and extensively drug resistant (XDR) M. tuberculosis and HIV-TB co-infection that contributed to the increased rates of TB [3]. Despite the limitations of long-term oral chemotherapy and inadequate compliance to the current treatment regimen, only two new anti-TB drugs, bedaquiline and delamanid have been successfully approved in the past 50 years [4, 5].

Therefore, the repurposing of existing drugs into advanced drug delivery systems could be an effective approach to improve the existing arsenal of TB treatments.

In recent years, the role of bacterial and macrophage efflux pumps in promoting M. tuberculosis drug resistance has gained significant attention. First, M. tuberculosis utilise these pumps to extrude ‘toxic’ compounds including anti-TB agents from the interior of the bacterial cell into their extracellular environment. Overexpression of these pumps leads to reduced drug concentration in the bacterial cytoplasm thus enabling intrinsic and acquired drug resistance in

M. tuberculosis [6-8]. For instance, Rodrigues et al. demonstrated that the overexpression of bacterial efflux pumps alone can result in isoniazid-resistance in M. tuberculosis without requirement of genetic mutations [8]. Second, M. tuberculosis within the infected macrophages are protected from high antibiotic concentrations because the macrophage efflux pumps also extrude anti-TB agents to provide a low intramacrophage drug concentration [7, 9]. For example, Adams et al. revealed that exposure to non-lethal concentration of rifampicin within

the macrophage environment transcriptionally induced M. tuberculosis efflux pump Rv1258 expression which allowed the bacteria to become resistant [10]. These mechanisms were also further confirmed in macrophages infected with several drug sensitive and resistant M. tuberculosis strains in which the rapid development of tolerance to various anti-TB agents, including isoniazid, moxifloxacin, linezolid, PA-824, and bedaquliline was established [11].

These bacterial and macrophage efflux pump-mediated resistant mechanisms may prolong the treatment duration for TB and requires a multi-drug regime for cure [12].

Adjunctive therapy with a drug that inhibits these efflux pumps could be a promising treatment approach against TB. The efflux pump inhibitor could: (1) directly block the efflux system of

M. tuberculosis and reduce the development of bacterial phenotypic resistance [7, 13]; (2) block the efflux pump of infected macrophages thereby reducing intracellular drug level [7,

13]; and (3) inhibit the calcium and potassium efflux pump of the phagolysosome, that causes acidification of the phagolysosome to activate hydrolytic enzymes to degrade the intracellular

M. tuberculosis [14]. Hence, these three mechanisms of action have the potential to shorten the current TB oral drug regimens and reduce the emergence of new MDR and XDR-TB strains.

Verapamil, a calcium-channel blocker is a most potent mycobacterial efflux pump inhibitor both in vitro and in vivo TB models [10, 15-18]. Verapamil has been approved by Food and

Drug Administration (FDA) since 1982 for oral treatment of cardiovascular disorders [19]. In vitro studies have highlighted the capacity of verapamil to significantly improve the anti- mycobacterial activity of many drugs including rifampicin, streptomycin, isoniazid, ofloxacin, bedaquiline, and clofazimine [11, 13, 17, 18, 20]. Verapamil also reduces drug tolerance in the macrophage environment and amplifies the intracellular activity of isoniazid and rifampicin against M. tuberculosis by enhancing the accumulation of these drugs inside the macrophage

[10, 11]. Furthermore, an in vivo study by Louw et al. demonstrated that verapamil can restore

M. tuberculosis susceptibility to first-line drugs and prevent the development of further drug resistance of mice infected with MDR-TB isolate [21]. Their study also showed that the incorporation of verapamil in the treatment of first-line anti-TB drugs significantly reduced the bacillary load in the lungs compared to the anti-TB drugs alone. Similarly, Gupta et al. demonstrated accelerated bacillary clearance and reduced relapse rates of standard anti-TB therapy in M. tuberculosis infected mice when verapamil was co-administered [16]. These findings are promising, but clinical studies investigating verapamil as an adjunctive anti-TB agent are limited.

The under-utilisation of verapamil may be due to its limitations when administered orally with rifampicin; a key first-line oral anti-TB drug that induces verapamil-metabolising enzyme

CYP3A4 [16, 22-24]. Thus, concomitant oral administration of these two drugs leads to variable pharmacokinetics of verapamil [16, 22, 23] and severely reduces its clinical efficacy

[25]. There is also a significant variation in bioavailability of verapamil even in rifampicin naïve subjects after oral administration due to high plasma protein binding of verapamil nearly

90 % [26, 27]. These metabolic and protein binding limitations can be overcome by inhaled delivery of verapamil in combination with rifamycin. Apart from avoiding first-pass metabolism by the liver, pulmonary delivery offers additional advantages. First, targeted delivery to the primary site of infection may increase local therapeutic concentration and anti-

TB activity. Second, high doses of efflux inhibitors at the site of infection could also enhance the activity of other orally-administered anti-TB drugs, including the accumulation into infected macrophages. This dual mechanism has the potential to shorten the treatment duration and reduce the side effects associated with prolonged drug usage.

The current study describes the manufacture of a combination inhalable dry powder consisting of verapamil and rifapentine. Rifapentine was selected as it is a first-line anti-TB drug similar in action to rifampicin, but has a longer pulmonary half-life and stronger anti-mycobacterial activity [28, 29]. The inhalable formulation of verapamil and rifapentine was produced at a drug mass ratio of 1:3, respectively based upon current recommended oral dosages [30, 31].

The physiochemical, in vitro aerosol performance and storage stability, and in vitro microbiological activities including pulmonary cytotoxicity are reported.

2. MATERIALS AND METHODS

Materials

Verapamil hydrochloride, L-leucine, potassium dihydrogen orthophosphate, dimethyl sulfoxide (DMSO), and phosphate-buffered saline (PBS) were obtained from Sigma-Aldrich,

Castle Hill, Australia. Rifapentine was from Hangzhou ICH Imp & Exp Company Ltd.,

Hangzhou, China. Acetonitrile, methanol, tetrahydrofuran, and orthophosphoric acid were purchased from Ajax Finechem Pty Ltd., Taren Point, Australia.

Bacterial strain and cell culture

M. tuberculosis H37Ra (ATCC 25177), M. tuberculosis H37Rv (ATCC 27294), human monocytic cell line (THP-1), and human alveolar basal adenocarcinoma epithelial cell line

(A549) were purchased from American Type Culture Collection (ATCC) (Manassas, VA). The

M. tuberculosis strains maintained in Middlebrook 7H9 media (enriched with 10 % albumin, dextrose and catalase (ADC) supplement) with 20 % glycerol stock at -80°C. The THP-1 and

A549 cells were grown in complete RPMI-1640 media (RPMI-1640 medium, 0.05 mM 2- mercaptoethanol, 10 % fetal bovine serum) at 37°C in 5 % CO2.

Formulation design

The single drug formulation of verapamil (Vrp_Leu) was produced by spray drying a solution containing 2.7 mg/mL verapamil hydrochloride and 0.7 mg/mL L-leucine (equivalent to 20 %, w/w) in distilled water. For the rifapentine only (Rpt) powder, a rifapentine crystalline

suspension was prepared as previously described [32]. Briefly, a rifapentine solution at concentration of 4 mg/ml was prepared in a co-solvent consisting of acetone:water (1:1, v/v).

The solution was then heated to 67°C for evaporation of acetone to form a crystalline suspension. This suspension was homogenised at 10 000 rpm for 2 minutes (Silverson L4RT

High Shear Mixer, Silverson, Chesham, UK) and then spray dried. For the combination formulation (Vrp_Rpt_Leu), verapamil hydrochloride (at 1:3; verapamil:rifapentine, w/w) and

L-leucine (20 %, w/w) were solubilised in the homogenised rifapentine suspension and the mixture was spray dried.

The spray drying process was carried out using a B-290 Mini Spray Dryer (Buchi Labortechnik

AG, Flawil, Switzerland) in an open-loop configuration using the following settings: inlet temperature of 60°C, atomiser of 60 mm (approximate 742 L/h), aspirator of 100 % (40 m3/h), and feed rate of 5 % (equivalent to 2 mL/min). A spray dried L-leucine powder was also produced through spray drying a L-leucine aqueous solution at a concentration of 0.7 mg/mL using the same parameters.

Physico-chemical characterisation

Particle morphology

Particle morphology was imaged using scanning electron microscopy (SEM) (Zeiss Ultra Plus scanning microscope, Carl Zeiss, Oberkochen, Germany) at 5 kV acceleration voltages.

Powders were deposited onto carbon tape and sputter coated with 15 nm of gold using a K550X sputter coater (Quorum Emitech, Kent, UK).

Particle sizing

Equivalent volume diameter was determined using laser diffraction coupled with a Scirocco dry powder module (Mastersizer 2000, Malvern Instruments, Worcestershire, UK). The dry powder was dispersed at a pressure of 3.5 bar. All the measurements were performed in triplicate. The refractive index (RI) was set at 1.54, 1.63 and 1.60 (average RI of verapamil and rifapentine at ratio 1:3) for Vrp_Leu, Rpt and the Vrp_Rpt_Leu, respectively.

X-ray power diffractometry

The crystalline property of the dry powders was analysed using x-ray powder diffractometry

(XRPD) (D500; Siemens, Karlsruhe, Germany). The powder was spread into the cavity of the

XRPD glass slide, and analysis was performed using Cu-Kα radiation at a generator voltage of

40 kV and a current of 33 mA. The data were collected by the 2θ scan method with a scan rate of 0.04°/sec from 5-40°.

Thermal analysis

Differential scanning electron calorimetry (DSC) and thermogravimetric analysis (TGA)

(Mettler Toledo, Melbourne, Australia) were used to study the thermal profile of the dry powders. For DSC, 5 mg of each powder was evenly spread into an aluminium crucible which was then crimped with a perforated lid and heated from 30 to 300°C at 10°C/min under a continuous flow of nitrogen gas. For TGA, 5 mg of each powder was weighed into an alumina crucible and then gradually heated at 10°C/min from 30 to 300°C under a dynamic nitrogen

gas flow. Both sets of thermal data were recorded and analysed through STARe software

V.9.0x (Mettler Toledo, Greifensee, Switzerland).

Dynamic vapour sorption

The moisture sorption behaviour of the powders was assessed using a dynamic vapour sorption

(DVS) instrument (DVS-1, Surface Measurement Systems, London, UK) operated at 25°C with continuous nitrogen gas flow. The powders were exposed to a dual moisture ramping cycle of 0 to 90 % with 10 % increments of RH. The mass changes were recorded over time and equilibrium moisture content determined by dm/dt of 0.02 % per minute.

Dissolution profile

The dissolution profile of the dry powders was studied using a Franz-cell dissolution system coupled with a heating stirrer station (Perm Gear Inc., Bethlehem, USA). The method was adapted from Chan et al. [33]. Each cell reservoir was filled with 23 mL of PBS (pH 7.4) with

2 % ascorbic acid as an antioxidant, to form a positive liquid meniscus. Magnetic stirrers were used to provide solute uniformity within the reservoirs. The cells were connected to a water bath to maintain the internal liquid temperature at 37°C.

Prior to dissolution, a capsule was filled with 20 mg of powder and actuated (Osmohaler®

(Plastiape, Italy) dry powder inhaler at an airflow of 100 L/min for 4 secs) into a next generation impactor (NGI) containing cellulose filter papers (Whatman® qualitative filter paper Grade 5) in stages 3 and 4 that has cut-off diameter of 2.2 and 1.3 µm, respectively. The filters holding the deposited powder were then placed in between the buffer meniscus and the top

compartment of the Franz-cell reservoir. Sampling (500 µL) was performed at predetermined time points (15, 30, 45, 60, 90, 120, 180, 240, and 300 min) and replaced with an equivalent amount of fresh buffer. At the final time point (5 h), the filter was removed and washed thoroughly with 3 mL of fresh buffer solution.

Drug quantification was performed using a validated high-performance liquid chromatography

(HPLC) method [32]. A µBondapakTM C18 (3.9 × 300 mm2) column (Waters, Milford,

Massachusetts) was coupled with a HPLC (Model LC-20, Shimadzu, Kyoto, Japan) operated at a detection wavelength of 250 nm (SPD-20A UV/VIS detector, Shimadzu) and an injection volume of 20 µL. The mobile phase was a 50 mM phosphate buffer (pH

3.0):acetonitrile:tetrahydrofuran at a ratio of 53:42:5 (%, v/v). The mobile phase was run at an isocratic flow rate of 1.0 mL/min. Calibration curves with R2 values of 1.0 for verapamil and rifapentine were obtained using standard solutions with concentrations ranging from 0.0005 to

0.1 mg/mL. The data were expressed as a percentage of drug mass in the buffer solution relative to the total mass recovery. Each formulation was tested in triplicate.

In vitro aerosol performance

In vitro aerosol performance of the spray dried powders was analysed using the Pharmacopeia method of the multi-stage liquid impinger (MSLI) (Apparatus C, British Pharmacopeia 2012,

Copley, Nottingham, UK) connected to a USP throat with a silicone mouthpiece adapter [32].

The rinsing solvent consisted of ascorbic acid (0.2 mg/mL) in methanol:water at 3:1. The MSLI was filled with 20 mL of rinsing solvent in stages 1 to 4 and fitted with a glass fibre filter in stage 5. Size 3 hydroxypropyl methylcellulose capsules (Capsugel, Greenwood) were filled with 10 ± 0.5 mg of powder. A capsule was punctured in an Osmohaler® (Plastiape, Italy), and

actuated for 4 secs at 100 L/min (calibrated using airflow and flow controller (TSI 3063 airflow meter, TSI Instruments Ltd., Buckinghamshire, UK and Flutus Air, Novi Systems Ltd.,

Dulverton, UK, respectively)). Following the dispersion, the components were washed with rinsing solvent as follows: 5 mL for the adaptor, and 10 mL each for the capsule, device, throat and filter. Dispersions were performed at 20 ± 3°C and a relative humidity (RH) of 50 ± 5 %.

Each formulation was tested in triplicate and the drug mass was quantified using HPLC as mentioned previously. The lower cut-off diameters for stages 1-4 of the MSLI at 100 L/min are 10.4, 4.9, 2.4, and 1.2 µm, respectively.

The total fine particle fraction (FPFtotal) was expressed as a mass percentage of particles with an aerodynamic diameter less than 5 µm relative to the total recovered drug mass. The emitted fine particle fraction (FPFemitted) was calculated as a mass percentage of particles with an aerodynamic diameter less than 5 µm relative to the total recovered drug mass excluding the capsule and device. Mass median aerodynamic diameter (MMAD) and geometric standard deviation (GSD) were calculated from the log-probability plot of the MSLI data.

Storage stability studies

Storage stability of Vrp_Leu and Vrp_Rpt_Leu was studied at 0 % and 60 % RH at 20 ± 3°C for 1 and 3 month intervals. These study conditions were chosen because the powders absorb significant amount of water at a RH above 60 % at 25°C as shown by the DVS analysis. The stability of Rpt formulation has been reported in a previous study [32]. The spray dried powders were stored in loosely capped scintillation vials in a desiccator at 20 ± 3°C and protected from light. To maintain RH close to zero, the desiccator contained gel desiccant beads and was under vacuum. For a 60 % RH and 20 ± 3°C environment, the powders were stored in a climate

controlled cabinet. At predetermined time points a known quantity of powder was dissolved in

MSLI rinsing solvent and quantified using HPLC to study the chemical stability of verapamil and rifapentine. The drug content was expressed as a mass percentage of each drug. Aerosol performance was also assessed on samples at these time points using the aforementioned methods.

In vitro microbiological and cytotoxicity studies

Resazurin assay

The resazurin assay was performed in 96-well plates. The antibiotic powders were dissolved in DMSO and serially diluted to the desired concentration in Middlebrook 7H9 media supplemented with 10 % ADC supplement, 0.05 % glycerol, 0.05 % Tween-80 and 1 % tryptone. Subsequently, mid-exponential phase (OD600 0.4 - 0.8) M. tuberculosis H37Ra or

M. tuberculosis H37Rv cultures were added and incubated for 5 days in a humidified incubator at 37°C. After the incubation, a mixture of 0.02 % resazurin solution (30 µL) and 20 % Tween-

80 (12.5 µL) was added to each well, then further incubated for 48 h. Fluorescence was measured using a BMG Labtech Polarstar Omega (BMG Labtech, Ortenburg, Germany) at an excitation wavelength of 530 nm and emission wavelength of 590 nm. Growth percentage was determined in comparison to the growth of the untreated negative control after subtracting the background fluorescence. Minimum inhibitory concentration (MIC90) was measured as the lowest concentration that showed 90 % growth inhibition compared to the untreated cells.

Intracellular inhibition assay

THP-1 cells were grown in complete RPMI-1640 media (RPMI-1640 medium, 0.05 mM 2- mercaptoethanol, 10 % fetal bovine serum) at 37°C in 5 % CO2. The media containing 50 ng/mL of phorbol 12-myristate 13-acetate and 1 × 105 cells/well were added to 96-well plates and incubated for 48 h for the cells to adhere and differentiate. The cells were then infected with M. tuberculosis H37Ra at a multiplicity of infection (MOI) of 5. After 4 h of infection, cells were washed 3 times with PBS, replaced with fresh media and further incubated overnight for the cells to recover. The verapamil and rifapentine in combination with various concentrations were dissolved in DMSO and then further diluted in complete RPMI-1640 media to achieve the desired concentration. Diluted solutions were added and incubated for 48 h. To determine the intracellular bacterial load, cells were lysed and bacterial number (colony forming units, cfu) were determined after plating onto Middlebrook 7H11 (Bacto Laboratories,

Mt. Pritchard, NSW, Australia) supplemented with 10 % oleic acid, albumin, dextrose and catalase (OADC) enrichment, and 0.5 % glycerol using standard procedures [34].

Cytotoxicity assay

The toxicity of the combination formulation and its individual components were tested on THP-

1 and A549 cell lines. A concentration of 5 × 105 cells/well was seeded to a 96-wells plate and incubated for 48 h to allow cell adherence and differentiation. The drug powders were dissolved in DMSO and diluted in complete RPMI-1640 media. Drugs (concentrations ranging from

0.00196 - 1.0 mg/mL) were added into the wells in two-fold dilutions and incubated for 4 days.

On day 4, 10 µL of 0.05 % (w/v) resazurin was added and incubated for 4 h. Fluorescence readings were then taken at a wavelength of 590 nm following excitation at 544 nm using the

FLUOstar Omega Microplate Reader (BMG Labtech, Ortenburg/Germany). The half maximal inhibitory concentration (IC50) was determined through comparison to the untreated cells.

Statistics

All data were expressed as mean ± standard deviation with n=3. The statistical differences were analysed using Microsoft Excel, t-test analysis of variance (Microsoft Office Excell 2007) and p values < 0.05 were considered statistically significant.

3. RESULTS

Physico-chemical characteristics

Particle morphology

SEM analysis on the physical morphology of various powders is presented in Figure 1. The raw verapamil was irregularly shaped, while processed Vrp_Leu was relatively spherical with a highly corrugated surface. The rifapentine (Rpt) particles consisted of needle-like structures.

A mixture of nearly spherical and elongated particles was observed in the combination formulation. The geometric size of verapamil particles were in the range of 1 - 3 µm while the rifapentine crystals were about 10  2  2 µm of length  width  height.

a b

c d

Figure 1. Scanning electron micrographs of powder samples: (a) raw verapamil, (b) Rpt, (c)

Vrp_Leu, and (d) Vrp_Rpt_Leu.

Particle sizing

Equivalent volume median diameter (D50) of the raw verapamil powder was about 7.39 ± 1.70

µm, which was reduced to 1.42 ± 0.02 µm after being solubilised and spray dried with L- leucine in Vrp_Leu (Table 1). The D50 of the Rpt (1.41 ± 0.06 µm) was similar to Vrp_Leu. A slight increase in size was observed for the combined Vrp_Rpt_Leu particles (1.61 ± 0.03 µm) and the particles were within the inhalable range of less than 5 µm.

Table 1. Particle size distribution of the formulations (mean ± SD, n=3)

D10 (µm) D50 (µm) D90 (µm)

Raw verapamil 2.27 ± 0.04 7.39 ± 1.70 24.37 ± 0.04

Rpt 0.74 ± 0.04 1.42 ± 0.02 2.72 ± 0.08

Vrp_Leu 0.80 ± 0.03 1.41 ± 0.06 2.66 ± 0.09

Vrp_Rpt_Leu 0.91 ± 0.03 1.61 ± 0.03 2.89 ± 0.05

X-ray powder diffractometry

Raw verapamil was shown to be crystalline as numerous diffraction peaks were observed, notably at 10.59°, 14.45°, 17.07°, 18.1°, 18.84°, 20.29°, 21.32°, 23.06°, 23.75°, and 26.29°

(Figure 2), similar to those reported in the literature [35]. The Vrp_Leu showed a typical “halo” pattern indicating the powder is largely amorphous. The Rpt powder demonstrated crystallinity with peaks at 5.68°, 6.52°, 7.56°, 8.48°, 9.08°, 10.04°, 13.16°, 15.68°, 17.8°, 20.76°, and

23.52°. The Vrp_Rpt_Leu powder also demonstrated crystalline peaks in similar regions as in

Rpt but at lower intensities. Crystalline peaks were also observed in the spray dried L-leucine.

Vrp_Rpt_Leu

Vrp_Leu Intensity

Rpt

Raw verapamil

Spray dried L-leucine

2 7 12 17 22 27 32 37 42 2θ (Degrees)

Figure 2. X-ray diffractograms of various powder formulations. Crystalline peaks were demonstrated by Vrp_Rpt_Leu, Rpt, raw verapamil and spray dried L-leucine. Vrp_Leu was largely amorphous.

Thermal analysis

100

Thermal properties of the Vrp_Rpt_Leu were compared to their individual components as shown in Figure 3. The raw verapamil exhibited a broad endotherm from 25°C to 130°C followed by a sharp endotherm peak at 146.65°C. In contrast, Vrp_Leu showed a broad exotherm from approximately 80°C to 225°C and an endotherm peak at 246.93°C. Slight enthalpy changes were observed in the DSC curves of Rpt and Vrp_Rpt_Leu from 25°C to

160°C, followed by an endothermic peak at 174.90°C and 174.84°C, respectively. The spray dried L-leucine showed an extended endotherm between 25°C and 275°C with a melting point of 293°C.

Spray dried L-leucine

Vrp_Rpt_Leu

Rpt

Vrp_Leu

Raw verapamil Exothermic → Exothermic

30 60 90 120 150 180 210 240 270 300 Temperature (°C)

Figure 3. DSC thermograms of the spray dried combination formulation and its individual components. The samples were heated from 30°C to 300°C under a dynamic flow of nitrogen gas.

101

Figure 4 shows the TGA dynamic curves of the powders. Almost complete water evaporation during the spray drying was observed because the measured weight loss was very low (< 0.5

%) for all samples after heating from 25°C to 120°C. In addition, raw verapamil and Vrp_Leu begin to degrade at around 225°C, while Rpt and Vrp_Rpt_Leu at 204°C and 162°C, respectively. Decomposition of spray dried L-leucine began after 240°C.

100

80 Raw verapamil Rpt

60 Vrp_Leu Vrp_Rpt_Leu

Weight (%) Weight Spray dried L-leucine 40

0 30 60 90 120 150 180 210 240 270 300 Temperature (°C)

Figure 4. TGA profiles of the spray dried formulation and its individual components. The samples were heated from 30°C to 300°C under a dynamic flow of nitrogen gas.

102

Dynamic vapour sorption

The dynamic water sorption behavior of Vrp_Leu, Rpt and Vrp_Rpt_Leu showed the formulations experienced a maximum mass change of 18.05 %, 13.70 % and 10.30 % (w/w), respectively at 90 % RH (Figure 5). All formulations exhibited a similar reversible moisture sorption trend at both the first and second moisture sorptions with no moisture-induced recrystallisation. A double sigmoid profile was observed during both moisture sorption and desorption cycles of Rpt; whereas only a slight sigmoid profile was noticed in desorption cycles

Vrp_Rpt_Leu which appeared after 60 % RH.

103

a Cycle 1 Sorp Cycle 1 Desorp Cycle 2 Sorp Cycle 2 Desorp 20

5 Change In Mass(%)

0 0 20 40 60 80 100 RH (%)

b Cycle 1 Sorp Cycle 1 Desorp Cycle 2 Sorp Cycle 2 Desorp 14 12 10 8 6 4

Change In Mass(%) 2 0 0 20 40 60 80 100 RH (%)

c Cycle 1 Sorp Cycle 1 Desorp

12 Cycle 2 Sorp Cycle 2 Desorp

Change In Mass(%) 2

0 0 20 40 60 80 100 RH (%)

Figure 5. Moisture sorption isotherms for (a) Vrp_Leu, (b) Rpt and (c) Vrp_Rpt_Leu.

Humidity ramped from 0 % to 90 % RH for 2-cycles.

104

Dissolution profile

Figure 6 shows the dissolution properties of the Vrp_Leu, Rpt and Vrp_Rpt_Leu. Verapamil, a hydrophilic drug, reached its maximal rate of dissolution within an hour in both Vrp_Leu and

Vrp_Rpt_Leu. However, the release rate of rifapentine was reduced by 20 % in Vrp_Rpt_Leu compared to Rpt.

100

40 Verapamil in Vrp_Rpt_Leu Rifapentine in Vrp_Rpt_Leu 20 Vrp_Leu Rpt Cumulative Cumulative amount (% of drug recovery) 0 0 40 80 120 160 200 240 280 Time (min)

Figure 6. Dissolution profile of Vrp_Leu, Rpt and Vrp_Rpt_Leu using Franz-cell dissolution method over a period of 5 h (mean ± SD, n=3).

105

In vitro aerosol performance

The FPFtotal, FPFemitted, MMAD, and GSD of each formulation are displayed in Table 2. Figure

7 presents the drug deposition of Vrp_Leu, Rpt and Vrp_Rpt_Leu powders after MSLI dispersion.

The MMAD and GSD of Vrp_Leu were 2.8 µm and 1.6, respectively, which resulted in an

FPFtotal of 67.3 ± 1.2 %. The FPFtotal of verapamil was increased to 77.4 ± 1.1 % in

Vrp_Rpt_Leu, which corresponds to MMAD and GSD of 2.10 - 2.16 µm and 1.65 - 1.66, respectively. In both formulations, a higher percentage of verapamil was deposited in stage 3 and 4 compared to the other stages of the impinger.

The FPFtotal and FPFemitted of Rpt were 83.9 ± 0.6 % and 93.9 ± 0.3 %, respectively. Notably,

Rpt also consists of approximately 24.5 ± 1.0 % of particles aerodynamically less than 1.32

µm, demonstrated by the drug deposited in the filter stage. Rifapentine in Vrp_Rpt_Leu showed

FPFtotal and FPFemitted that was significantly reduced to 71.5 ± 2.0 % and 76.8 ± 2.2 %, respectively (p < 0.05).

More importantly, for the combination formulation no significant differences were observed in the deposition percentage between the two drugs in various stages of the MSLI (p > 0.05), indicating uniform distribution of verapamil and rifapentine.

106

Table 2. FPFtotal, FPFemitted, MMAD, and GSD for Vrp_Leu, Rpt and Vrp_Rpt_Leu over 3 months in 0 % RH (mean ± SD, n=3)

Vrp_Leu 0 months 1 month 3 months

FPFtotal (%) 67.3 ± 1.2 68.6 ± 0.7 65.2 ± 2.9

FPFemitted (%) 73.6 ± 2.0 74.2 ± 0.6 71.1 ± 2.9

MMAD (µm) 2.87, 2.86, 2.80 2.93, 2.94, 2.58 2.86, 2.71, 2.94

GSD 1.60, 1.58, 1.59 1.61, 1.61, 1.60 1.58, 1.56, 1.58

Rpt 0 months

FPFtotal (%) 83.9 ± 0.6 - -

FPFemitted (%) 93.9 ± 0.3 - -

MMAD (µm) 1.76, 1.81, 1.79 - -

GSD 1.69, 1.70, 1.69 - -

Vrp_Rpt_Leu 0 months 1 month 3 months

verapamil rifapentine verapamil rifapentine verapamil rifapentine

FPFtotal (%) 77.4 ± 1.1 71.5 ± 2.0 77.4 ± 1.1 71.8 ± 4.2 76.7 ± 1.5 69.6 ± 1.8

FPFemitted (%) 85.7 ± 1.1 76.8 ± 2.2 85.7 ± 1.1 78.2 ± 4.0 84.8 ± 1.2 77.8 ± 1.8

MMAD (µm) 2.11, 2.29, 2.29, 2.21, 2.39, 2.35, 2.10, 2.35, 2.33,

2.16, 2.10 2.19 2.19, 2.16 2.34 2.17, 2.19 2.32

GSD 1.66, 1.66,1.64, 1.57, 1.64, 1.63, 1.64, 1.65, 1.63,

1.65, 1.65 1.66 1.61, 1.56 1.62 1.62, 1.60 1.62

*MMAD and GSD from each replicate

107

50 a Vrp_Leu 40 Rpt

Recovery mass Recoverymass (%) 10

50 b Verapamil in Vrp_Rpt_Leu 40 Rifapentine in Vrp_Rpt_Leu

10 Recovered mass Recovered (%)mass

Figure 7. In vitro aerosol deposition profile for (a) Vrp_Leu and Rpt, and (b) Vrp_Rpt_Leu following MSLI dispersion at a flow rate of 100 L/min at 0 months using Osmohaler® (mean

± SD, n=3).

108

Storage stability studies

Figure 8 shows the SEM images of Vrp_Leu and Vrp_Rpt_Leu exposed to 60 % RH for a month. The particles in both formulations were observed to be fused into a non-dipersible solid aggregate. On the other hand, intact morphology (as in Figure 1) was observed in the powders stored at 0 % RH. Therefore, further characterisation was only carried out using powders stored at 0 % RH.

Table 3 shows the theoretical and actual drug content of verapamil and rifapentine in

Vrp_Rpt_Leu stored at 0 % RH over 3 months. The composition of the individual drugs remained unchanged after 3 months of storage (p > 0.05). In regard to their aerosol performance, no significant changes were seen in their FPFtotal, FPFemitted, MMAD and GSD as shown in Table 2 (p > 0.05). The aerosol deposition profiles of Vrp_Leu and Vrp_Rpt_Leu stored at 0 % RH up to 3 months (Figure 9) did not demonstrate any significant difference from their initial MSLI profile (p > 0.05).

a b

Figure 8. Scanning electron microscopy observation of (a) Vrp_Leu and (b) Vrp_Rpt_Leu exposed to 60 % RH for a month.

109

Table 3. Theoretical versus actual drug content of Vrp_Rpt_Leu over 3 months of storage in

0 % RH at 20°C (mean ± SD, n=3)

Duration of storage Drug content (%)

Verapamil Rifapentine

Theoretical 25 75

0 month 27.7 ± 1.5 72.3 ± 2.0

1 month 28.5 ± 3.2 71.5 ± 3.0

3 months 26.0 ± 2.5 73.0 ± 2.5

110

50 a Vrp_Leu 40 Verapamil in Vrp_Rpt_Leu Rifapentine in Vrp_Rpt_Leu 30

Recovered mass Recovered (%)mass 10

50 b Vrp_Leu 40 Verapamil in Vrp_Rpt_Leu Rifapentine in Vrp_Rpt_Leu 30

Recovered mass Recovered (%)mass 10

Figure 9. In vitro aerosol deposition profile for powders stored at 20 ± 3ᵒC in 0 % RH for (a)

1 month and (b) 3 months dispersed at a flow rate of 100 L/min via the Osmohaler® (mean ±

SD, n=3).

111

In vitro microbiological and cytotoxicity studies

Resazurin assay

The solubilised Rpt showed high degree of activity against M. tuberculosis H37Ra (MIC90

0.625 ng/mL) and M. tuberculosis H37Rv (MIC90 5.0 ng/mL). By contrast, the solubilised

Vrp_Leu has no anti-mycobacterial activity against both strains. The solubilised Vrp_Rpt_Leu demonstrated equivalent activity against M. tuberculosis H37Ra (MIC90 0.625 ng/mL) and M. tuberculosis H37Rv (MIC90 5.0 ng/mL) as Rpt. Furthermore, when the ratio of verapamil to rifapentine was increased, a decrease in the anti-mycobacterial activity was observed. The excipient, L-leucine did not inhibit mycobacterial growth at the concentrations used in the combinations of Vrp_Leu and Vrp_Rpt_Leu.

112

Table 4. Minimum inhibitory concentration (MIC90) against M. tuberculosis H37Ra and M. tuberculosis H37RV of the three drug formulations and the individual compounds using the resazurin assay

Compound MIC90 (µg/mL)*

M. tuberculosis M. tuberculosis

H37Ra H37Rv

Vrp_Leu 250 500

Rpt 0.000625 0.005

Spray dried L-leucine > 62.5 > 1000

Vrp_Rpt_Leu 0.000625 0.005

(verapamil:rifapentine,1:3)

Verapamil : rifapentine, 1:1 0.00125 0.01

Verapamil : rifapentine, 3:1 0.005 0.01

Verapamil : rifapentine, 10:1 0.01 > 0.02

*Mean MIC90 of three independent experiments performed in triplicate.

113

Intracellular inhibition of M. tuberculosis H37Ra

In light of the efficacy of the solubilised Vrp_Rpt_Leu against M. tuberculosis, we investigated whether verapamil and rifapentine at a ratio of 1:3 could restrict the mycobacterial growth within the host THP-1 cell. This assay was only conducted with M. tuberculosis H37Ra due to biosafety constraints on using virulent M. tuberculosis H37Rv strain.

As shown in Figure 10, the combined verapamil and rifapentine showed a dose-dependent bactericidal activity on M. tuberculosis H37Ra growth over 2 days of infection. Furthermore, a statistically significant (p < 0.05) difference was observed between the individual drug treatments compared to the combination. For example on day 1, treatment with the combination

(verapamil 0.05 µg/ml + rifapentine 0.15 µg/ml) resulted in 35 % and 20 % increased bacterial killing compared to verapamil (0.05 µg/ml) or rifapentine (0.15 µg/ml) alone, respectively. A similar trend was also observed at the other two concentrations tested. These findings indicated that the intracellular activity of rifapentine was enhanced by the presence of verapamil.

114

a **

10 ** Millions

6 *

* * 4 **

2 Mycobacterialgrowth(CFU/mL)

0 No drug VRP (0.05) RPT (0.15) VRP/RPT VRP (0.5) RPT (1.5) VRP/RPT VRP (5) RPT (15) VRP/RPT (0.05/0.15) (0.5/1.5) (5/15) Concentration (µg/mL)

b ** 10

** Millions

8 ** **

4 **

2 Mycobacterialgrowth(CFU/mL) 0 No drug VRP (0.05) RPT (0.15) VRP/RPT VRP (0.5) RPT (1.5) VRP/RPT VRP (5) RPT (15) VRP/RPT (0.05/0.15) (0.5/1.5) (5/15) Concentration (µg/mL)

Figure 10. Bactericidal effects of verapamil (VRP) and rifapentine (RPT) at ratio of 1:3 on intracellular infection of M. tuberculosis H37Ra in THP-1 cell treated for (a) 24 h and (b) 48 h (mean ± SD, n=3). The vertical dotted lines separate the different concentrations tested.

Differences between treatments were evaluated by analysis of variance (*p < 0.05, **p < 0.01).

115

Cytotoxicity assay

The safety of the formulations was investigated by determining the cytotoxic half maximal inhibitory concentration (IC50) against THP-1 and A549 cell lines (Table 5). The IC50 of all solubilised antibiotic powders were in the range of 62.5 µg/mL and 250 µg/mL. L-leucine did not have any cytotoxic effect on either cell line.

Table 5. Half maximal inhibitory concentration (IC50) of the inhalable dry powder formulations on human monocytic (THP-1) and human alveolar basal carcinoma epithelial (A549) cell lines

Compound IC50 (µg/mL)

THP-1 A549

Vrp_Leu 125 250

Rpt 62.5 62.5

Vrp_Rpt_Leu 62.5 62.5

Spray dried L-leucine >1000 >1000

116

4. DISCUSSION

To the best of our knowledge, this is the first study on the development of inhalable formulation of efflux pump inhibitor, verapamil and rifapentine. The inhaled combination formulation is uniquely composed of corrugated amorphous verapamil attached to elongated crystalline rifapentine particles. This distinct combination showed a promising aerodynamic performance for both verapamil and rifapentine. When verapamil alone was spray dried with L-leucine

(Vrp_Leu), it had an FPFtotal above 65 % and the MMAD was 2.8 µm. The corrugated surface morphology of these amorphous particles reduces the interparticle cohesive forces which concomitantly lead to lower particle adhesion forces and superior FPF compared to smoother particles [36, 37]. The addition of L-leucine also reduces intermolecular interaction between the drugs for enhanced dispersion performance [38]. When verapamil and L-leucine were further combined with elongated rifapentine (Rpt) and spray dried (Vrp_Rpt_Leu), the FPFtotal of verapamil increased to more than 75 % and the MMAD was reduced to 2.1 µm (p < 0.05).

This was due to a larger deposition of the particles on stage 4 compared to stage 3 in Vrp_Leu, indicating that the elongated Rpt particles act as a carrier for the amorphous verapamil to enhance the powder dispersibility. A similar observation was also reported in an inhaled formulation that contained colistin and rifapentine [39]. On the other hand, lower deposition of rifapentine on the final filter stage (aerodynamically < 1.32 µm) reduced the respirable fraction of rifapentine in the combination powder reduced to 70 % (p < 0.05) compared to 83 % in the rifapentine alone formulation. However, this reduction may not be clinically significant as the particles likely to be submicron, hence exhaled due to negligible impaction and gravitational sedimentation in the lung airways [40]. More importantly, Vrp_Rpt_Leu also showed a uniform dose distribution of both drugs at their intended dose (p < 0.05) in stages 1 - 4 of the

MSLI, suggesting a potential homogeneous pulmonary deposition of both drugs. Therefore,

117

the Vrp_Rpt_Leu combination demonstrates suitability for pulmonary delivery at the intended dose with high lung deposition profile.

Another advantage of the Vrp_Rpt_Leu formulation was a slower in vitro dissolution rate of

Rpt particles. Raula et al. demonstrated a similarly slow dissolution rate for L-leucine coated salbutamol sulphate (produced via spray drying) compared to pure drug particles [41]. This phenomenon was explained by the relatively low water solubility of L-leucine (approximately

24 g/L) [41]. On the basis of their finding, it could be deduced that during the spray drying process L-leucine coated onto the surface of the Rpt and slows the dissolution of the particles.

In the context of TB treatment, the slow release of anti-TB drugs may allow less frequent dosing as well as longer retention of physical drug particles to enhance internalisation by macrophages [42].

The stability of the formulations was also addressed. A previous study has demonstrated that amorphous rifapentine was chemically unstable after a month of storage at ambient (60 % RH) and dry (0 % RH) conditions [32]. A unique crystalline form of rifapentine was therefore developed and shown to be chemically stable for up to 3 months of storage under the aforementioned conditions [32]. In contrast, spray dried verapamil (without L-leucine) was very hygroscopic leading to immediate caking after being spray dried at ambient RH (45 ± 5

%). In a recent Phase 1, single-dose study by Dharmadhikari et al. used inhalable capreomycin spray dried with 20 % L-leucine (w/w) with no severe adverse effects up to administration of

300 mg [43]. Hence, a similar proportion of L-leucine was used to stabilise verapamil alone formulation.

118

The DVS moisture sorption profile demonstrates the Vrp_Leu still absorbs a significant amount of water at elevated RH above 60 %. Nonetheless, the reversible water absorption and desorption pattern as indicated by the identical reversible moisture sorption trend at both first and second moisture sorption cycles suggesting absence of recrystallisation. This could be due to the presence of L-leucine, which has been reported to prevent water-induced recrystallisation of amorphous pyrazinamide in an inhalable antibiotic formulation consisting of rifapentine, moxifloxacin and pyrazinamide [44]. Furthermore, a similar reversible sorption profile was observed in Vrp_Rpt_Leu with an additional sigmoid profile after 60 % RH. The presence of a sigmoid profile is attributable to the crystalline rifapentine, which forms a monohydrate followed by a dehydrate [32].

Both Vrp_Leu and Vrp_Rpt_Leu can be stored at validated conditions for long-term stability.

Vrp_Leu and Vrp_Rpt_Leu powders stored for a month at 60 % RH absorbed a significant amount of water to become a non-dispersible cake. In contrast, Zhou et al. have shown that co- solvent spray drying hygroscopic colistin with hydrophobic rifampicin at 1:1 protected the powder from storage moisture. They proposed that this was due to the coating of the rifampicin onto the surface of the colistin [45]. Although, Vrp_Rpt_Leu consists of 75 % of hydrophobic rifapentine, it was crystallised first before the addition of verapamil prior to spray drying which resulted the drugs present as two distinctive forms. The water absorption of Vrp_Rpt_Leu during storage therefore, could be mainly attributed to the relatively hygroscopic verapamil particles. In contrast, the formulations stored for up to 3 months in a moisture-free (vacuum) condition retained their physical morphology when observed under SEM. No statistical difference was observed in the dispersion profile of these powders up to 3 months. Drug quantification analysis also indicated the individual components of the combination remained at the intended mass ratio after storage. The HPLC data revealed drug recovery of 95 – 105 %

119

and no extra peaks that would indicate oxidation or other unwanted chemical reactions. Thus, both Vrp_Leu and Vrp_Rpt_Leu should be in moisture protective packaging for long term storage.

Verapamil and rifapentine in combination did not show synergistic inhibitory activity against

M. tuberculosis in vitro. The solubilised Vrp_Rpt_Leu formulation was anticipated to show a lower MIC90 than the rifapentine alone formulation owing to efflux pump inhibition by verapamil. Unexpectedly, the MIC90 value did not decrease even with an increase in the proportion of verapamil. In contrast, Gupta et al. reported that the MIC of bedaquiline and clofazimine against the laboratory H37Rv strain and clinical isolates was reduced by 8 to 16 times in the presence of verapamil [17]. Another study demonstrated verapamil resulted at least a 4-fold reduction in the in vitro MIC of isoniazid against M. tuberculosis H37Rv [8].

Rodrigues et al., who studied the contribution of efflux mechanism on isoniazid resistance in

M. tuberculosis complex highlighted that activation of efflux pumps depends on the genotype of the strain tested rather than the drug alone [8].

Nevertheless, the killing of M. tuberculosis H37Ra within the macrophage host environment was enhanced by the combination of verapamil with rifapentine when compared to treatment with the individual drugs. This result could be due to verapamil that appears to inhibit the macrophage-induced M. tuberculosis efflux pumps, leading to the increased killing of intracellular mycobacteria [10, 17].

The cytotoxicity profiles were also studied to examine the tolerance of pulmonary cells to these formulations. As the drugs are shown to accumulate in macrophages, tolerability was assessed in THP-1 derived macrophages. The tolerability was also assessed on a lung epithelial cell line

120

due to the non-specific nature of pulmonary delivery. For both cell lines, solubilised

Vrp_Rpt_Leu demonstrated greater toxicity than the Vrp_Leu solution, but this was similar to

Rpt alone. This could be because the majority (75 %) of the formulation was rifapentine.

However, the cytotoxicity of solubilised Vrp_Rpt_Leu may underestimate the potential pulmonary toxicity of the formulation as the Rpt particles may not undergo rapid dissolution following lung deposition. In a previous study, Chan et al. reported the solubilised Rpt alone was well tolerated in vitro by the macrophage and epithelial cell line models up to 50 µM, compared 71.3 µM in the present study [32]. Additionally, pharmacokinetics analysis following a single pulmonary dose of Rpt at 20 mg/kg showed that intratracheal delivery can achieve Cmax of 321 µg/ml in the lungs with no adverse event in the animals during the treatment. This suggests that the pulmonary tolerability of rifapentine may be even higher.

Further pulmonary safety studies of Vrp_Rpt_Leu are warranted.

121

5. CONCLUSION

In this study, we formulated a novel inhalable verapamil and rifapentine combination dry powder that demonstrates excellent aerosol performance, enhanced activity against intracellular growth of M. tuberculosis in macrophages relative to the single drugs, and a promising in vitro cytotoxicity profile. Murine studies are underway to ascertain in vivo efficacy of this formulation.

122

ACKNOWLEDGEMENT

Thaigarajan Parumasivam is a recipient of Malaysian Government Scholarship. The authors acknowledge Australian Microscopy & Microanalysis Research Facility at the Australian

Centre for Microscopy and Microanalysis, The University of Sydney, for their scientific and technical assistance. The work was supported by the Australian Research Council’s Discovery

Project (DP120102778), the National Health and Medical Research Council Centre of

Research Excellence in Tuberculosis Control (APP1043225) and Project grant (APP104434), and the NSW Government Infrastructure Grant to the Centenary Institute.

123

REFERENCES

[1] World Health Organization, Global tuberculosis report 2014, in, http://www.who.int/tb/publications/global_report/gtbr14_main_text.pdf?ua=1, 2013. [2] C. Dye, S. Scheele, P. Dolin, V. Pathania, M.C. Raviglione, W.H.O.G.S. for the, P. Monitoring, Global burden of tuberculosis: estimated incidence, prevalence, and mortality by country, JAMA, 282 (1999) 677-686. [3] I. Monedero, J.A. Caminero, MDR-/XDR-TB management: what it was, current standards and what is ahead, Expert Rev Respir Med, 3 (2009) 133-145. [4] A.I. Zumla, S.H. Gillespie, M. Hoelscher, P.P.J. Philips, S.T. Cole, I. Abubakar, T.D. McHugh, M. Schito, M. Maeurer, A.J. Nunn, New antituberculosis drugs, regimens, and adjunct therapies: needs, advances, and future prospects, Lancet Infect Dis, 14 (2014) 327-340. [5] A.S. Xavier, M. Lakshmanan, Delamanid: A new armor in combating drug-resistant tuberculosis, J Pharmacol Pharmacother, 5 (2014) 222-224. [6] S. Ramón-García, C. Martín, C.J. Thompson, J.A. Aínsa, Role of the Mycobacterium tuberculosis P55 efflux pump in intrinsic drug resistance, oxidative stress responses, and growth, Antimicrob Agents Chemother, 53 (2009) 3675-3682. [7] J. Szumowski, K. Adams, P. Edelstein, L. Ramakrishnan, Antimicrobial efflux pumps and Mycobacterium tuberculosis drug tolerance: evolutionary considerations, in: J. Pieters, J.D. McKinney (Eds.) Pathogenesis of Mycobacterium tuberculosis and its interaction with the host organism, Springer Berlin Heidelberg, 2013, pp. 81-108. [8] L. Rodrigues, D. Machado, I. Couto, L. Amaral, M. Viveiros, Contribution of efflux activity to isoniazid resistance in the Mycobacterium tuberculosis complex, Infect Genet Evol, 12 (2012) 695-700. [9] C.X. Cao, S.C. Silverstein, H.C. Neu, T.H. Steinberg, J774 macrophages secrete antibiotics via organic anion transporters, J Infect Dis, 165 (1992) 322-328. [10] Kristin N. Adams, K. Takaki, Lynn E. Connolly, H. Wiedenhoft, K. Winglee, O. Humbert, Paul H. Edelstein, Christine L. Cosma, L. Ramakrishnan, Drug tolerance in replicating mycobacteria mediated by a macrophage-induced efflux mechanism, Cell, 145 (2011) 39-53. [11] K.N. Adams, J.D. Szumowski, L. Ramakrishnan, Verapamil, and its metabolite norverapamil, inhibit macrophage-induced, bacterial efflux pump-mediated tolerance to multiple anti-tubercular drugs, J Infect Dis, 210 (2014) 456-466. [12] R.D. Berg, L. Ramakrishnan, Insights into tuberculosis from the zebrafish model, Trends Mol Med, 18 (2012) 689-690. [13] G.E. Louw, R.M. Warren, N.C. Gey van Pittius, C.R.E. McEvoy, P.D. Van Helden, T.C. Victor, A balancing act: efflux/influx in mycobacterial drug resistance, Antimicrob Agents Chemother, 53 (2009) 3181-3189. [14] L. Amaral, J. Molnar, Why and how the old neuroleptic thioridazine cures the XDR-TB patient, Pharmaceuticals (Basel), 5 (2012) 1021-1031. [15] K. Singh, M. Kumar, E. Pavadai, K. Naran, D.F. Warner, P.G. Ruminski, K. Chibale, Synthesis of new verapamil analogues and their evaluation in combination with rifampicin against Mycobacterium tuberculosis and molecular docking studies in the binding site of efflux protein Rv1258c, Bioorg Med Chem Lett, 24 (2014) 2985-2990. 124

[16] S. Gupta, S. Tyagi, D.V. Almeida, M.C. Maiga, N.C. Ammerman, W.R. Bishai, Acceleration of tuberculosis treatment by adjunctive therapy with verapamil as an efflux inhibitor, Am J Respir Crit Care Med, 188 (2013) 600-607. [17] S. Gupta, K.A. Cohen, K. Winglee, M. Maiga, B. Diarra, W.R. Bishai, Efflux inhibition with verapamil potentiates bedaquiline in Mycobacterium tuberculosis, Antimicrob Agents Chemother, 58 (2014) 574-576. [18] D. Machado, I. Couto, J. Perdigão, L. Rodrigues, I. Portugal, P. Baptista, B. Veigas, L. Amaral, M. Viveiros, Contribution of efflux to the emergence of isoniazid and multidrug resistance in Mycobacterium tuberculosis, PLoS ONE, 7 (2012) e34538. [19] O. Ogbru, Verapamil, Calan, Verelan, Verelan PM, Isoptin, Isoptin SR, Covera-HS, in: J.W. Marks (Ed.), http://www.medicinenet.com/verapamil/article.htm, 2014. [20] F.S. Spies, P.E. Almeida da Silva, M.O. Ribeiro, M.L. Rossetti, A. Zaha, Identification of mutations related to streptomycin resistance in clinical isolates of Mycobacterium tuberculosis and possible involvement of efflux mechanism, Antimicrob Agents Chemother, 52 (2008) 2947-2949. [21] G.E. Louw, R.M. Warren, N.C. Gey van Pittius, R. Leon, A. Jimenez, R. Hernandez- Pando, C.R.E. McEvoy, M. Grobbelaar, M. Murray, P.D. van Helden, T.C. Victor, Rifampicin reduces susceptibility to ofloxacin in rifampicin-resistant Mycobacterium tuberculosis through efflux, Am J Respir Crit Care Med, 184 (2011) 269-276. [22] M.F. Fromm, D. Busse, H.K. Kroemer, M. Eichelbaum, Differential induction of prehepatic and hepatic metabolism of verapamil by rifampin, Hepatology, 24 (1996) 796-801. [23] J. Mooy, R. Böhm, M. van Baak, J. van Kemenade, A.v.d. Vet, K.H. Rahn, The influence of antituberculosis drugs on the plasma level of verapamil, Eur J Clin Pharmacol, 32 (1987) 107-109. [24] M.F. Fromm, K. Dilger, D. Busse, H.K. Kroemer, M. Eichelbaum, U. Klotz, Gut wall metabolism of verapamil in older people: effects of rifampicin-mediated enzyme induction, Br J Clin Pharmacol, 45 (1998) 247-255. [25] R.A. Barbarash, J.L. Bauman, J.H. Fischer, G.T. Kondos, R.L. Batenhorst, Near-total reduction in verapamil bioavailability by rifampin. electrocardiographic correlates, Anglais, 94 (1988) 954-959. [26] D.L. Keefe, Y.-G. Yee, R.E. Kates, Verapamil protein binding in patients and in normal subjects, Clin Pharmacol Ther, 29 (1981) 21-26. [27] R.E. Kates, Calcium antagonists. Pharmacokinetic properties., Drugs, 25 (1983) 113-124. [28] S.S. Munsiff, C. Kambili, S.D. Ahuja, Rifapentine for the treatment of pulmonary tuberculosis, Clin Infect Dis, 43 (2006) 1468-1475. [29] J.G.Y. Chan, A.S. Tyne, A. Pang, A.J. McLachlan, V. Perera, J.C.Y. Chan, W.J. Britton, H.K. Chan, C.C. Duke, P.M. Young, D. Traini, Murine pharmacokinetics of rifapentine delivered as an inhalable dry powder, Int J Antimicrob Agents, 43 (2014) 319-323. [30] I.M. Rosenthal, M. Zhang, K.N. Williams, C.A. Peloquin, S. Tyagi, A.A. Vernon, W.R. Bishai, R.E. Chaisson, J.H. Grosset, E.L. Nuermberger, Daily dosing of rifapentine cures tuberculosis in three months or less in the murine model, PLoS Med, 4 (2007) e344. [31] Micromedex, Verapamil (Oral Route), in: http://www.mayoclinic.org/drugs- supplements/verapamil-oral-route/proper-use/drg-20071728 (Ed.) 125

http://www.mayoclinic.org/drugs-supplements/verapamil-oral-route/proper-use/drg- 20071728, Thomson Healthcare Inc, http://www.mayoclinic.org/drugs- supplements/verapamil-oral-route/proper-use/drg-20071728, 2014, pp. http://www.mayoclinic.org/drugs-supplements/verapamil-oral-route/proper-use/drg- 20071728. [32] J.G.Y. Chan, C.C. Duke, H.X. Ong, J.C.Y. Chan, A.S. Tyne, H.-K. Chan, W.J. Britton, P.M. Young, D. Traini, A novel inhalable form of rifapentine, J Pharm Sci, 103 (2014) 1411- 1421. [33] J.G.Y. Chan, H.-K. Chan, C.A. Prestidge, J.A. Denman, P.M. Young, D. Traini, A novel dry powder inhalable formulation incorporating three first-line anti-tubercular antibiotics, Eur J Pharm Biopharm, 83 (2013) 285-292. [34] R. Pinto, B.M. Saunders, L.R. Camacho, W.J. Britton, B. Gicquel, J.A. Triccas, Mycobacterium tuberculosis defective in phthiocerol dimycocerosate translocation provides greater protective immunity against tuberculosis than the existing bacille Calmette-Guérin vaccine, J Infect Dis, 189 (2004) 105-112. [35] J. Sahoo, P.N. Murthy, S. Biswal, Manik, Formulation of sustained-release dosage form of verapamil hydrochloride by solid dispersion technique using Eudragit RLPO or Kollidon SR, AAPS PharmSciTech, 10 (2009) 27-33. [36] S. Adi, H. Adi, P. Tang, D. Traini, H.-k. Chan, P.M. Young, Micro-particle corrugation, adhesion and inhalation aerosol efficiency, Eur J Pharm Sci, 35 (2008) 12-18. [37] N. Chew, P. Tang, H.-K. Chan, J. Raper, How much particle surface corrugation is sufficient to improve aerosol performance of powders?, Pharmaceutical Research, (2005). [38] N.Y.K. Chew, B.Y. Shekunov, H.H.Y. Tong, A.H.L. Chow, C. Savage, J. Wu, H.-K. Chan, Effect of amino acids on the dispersion of disodium cromoglycate powders, J Pharm Sci, 94 (2005) 2289-2300. [39] Q. Zhou, S.-P. Sun, J.G.Y. Chan, P. Wang, N. Barraud, S.A. Rice, J. Wang, J. Li, H.-K. Chan, Novel inhaled combination powder containing amorphous colistin and crystalline rifapentine with enhanced antimicrobial activities against planktonic cells and biofilm of Pseudomonas aeruginosa for respiratory infections, Mol Pharm, 12 (2015) 2594-2603. [40] J. Zhang, L. Wu, H.-K. Chan, W. Watanabe, Formation, characterization, and fate of inhaled drug nanoparticles, Adv Drug Deliv Rev, 63 (2011) 441-455. [41] J. Raula, J. Seppälä, J. Malm, M. Karppinen, E. Kauppinen, Structure and dissolution of L-Leucine-coated salbutamol sulphate aerosol particles, AAPS PharmSciTech, 13 (2012) 707- 712. [42] P. Muttil, C. Wang, A. Hickey, Inhaled drug delivery for tuberculosis therapy, Pharm Res, 26 (2009) 2401-2416. [43] A.S. Dharmadhikari, M. Kabadi, B. Gerety, A.J. Hickey, P.B. Fourie, E. Nardell, Phase I, single-dose, dose-escalating study of inhaled dry powder capreomycin: a new approach to therapy of drug-resistant tuberculosis, Antimicrob Agents Chemother, 57 (2013) 2613-2619. [44] J. Chan, A. Tyne, A. Pang, H.-K. Chan, P. Young, W. Britton, C. Duke, D. Traini, A rifapentine-containing inhaled triple antibiotic formulation for rapid treatment of tubercular infection, Pharm Res, 31 (2013) 1239-1253.

126

[45] Q. Zhou, T. Gegenbach, J.A. Denman, H. Yu, J. Li, H.-K. Chan, Synergistic antibiotic combination powders of colistin and rifampicin provide high aerosolization efficiency and moisture protection, AAPS J, 16 (2014) 37-47.

127

CHAPTER 3

IN VITRO EVALUATION OF NOVEL INHALABLE DRY

POWDERS CONSISTING OF THIORIDAZINE AND

RIFAPENTINE FOR RAPID TUBERCULOSIS TREATMENT

This chapter has been published in European Journal of Pharmaceutics and Biopharmaceutics,

2016; 107: 205-214, with the title ‘In vitro evaluation of novel inhalable dry powders consisting of thioridazine and rifapentine for rapid tuberculosis treatment’. Authors: Parumasivam T,

Chan JGY, Pang A, Quan DH, Triccas JA, Britton WJ, Chan H-K.

128

ABSTRACT

Thioridazine is an orally administered antipsychotic drug with potential for treatment of drug- resistant tuberculosis (TB). However, drug-induced adverse cardiac effects have been reported when thioridazine was used at an efficacious oral dose of 200 mg/day to treat TB. Pulmonary delivery of thioridazine could be a rational approach to reduce dose-related side effects while enabling high drug concentrations at the primary site of infection. The present study compares in vitro aerosol performance, storage stability, and in vitro antimicrobial activity and cytotoxicity of two inhalable powders composed of thioridazine and a first-line anti-TB drug, rifapentine. Formulation 1 is a combination of amorphous thioridazine and crystalline rifapentine, while Formulation 2 consisted of both drugs as amorphous forms. Both thioridazine-rifapentine formulations were found suitable for inhalation with a total fine particle fraction (< 5 µm) of 68 - 76 %. The two powders had similar MIC90 to rifapentine alone, being 0.000625 µg/mL and 0.005 µg/ml against M. tuberculosis H37Ra and M. tuberculosis H37Rv, respectively. In contrast, thioridazine alone had a MIC90 of 12.5 µg/mL and 500 µg/mL against M. tuberculosis H37Ra and M. tuberculosis H37Rv, respectively, demonstrating no synergistic anti-TB activity. However, thioridazine and rifapentine in a ratio of 1:3 enhanced the killing of M. tuberculosis H37Ra within the human monocyte-derived macrophages (THP-1) compared to the single drug treatments. Both powders showed an acceptable half maximal inhibitory concentration (IC50) of 31.25 µg/mL on both THP-1 and human lung epithelial (A549) cells. However, Formulation 1 showed greater chemical stability than Formulation 2 after three months of storage under low humidity (vacuum) at 20 ± 3 °C.

In conclusion, we have demonstrated a novel inhalable powder consisted of amorphous thioridazine and crystalline rifapentine (Formulation 1) with a good aerosol performance, potent anti-TB activity and storage stability, which deserves further in vivo investigations.

129

1. INTRODUCTION

Tuberculosis (TB) is predominantly a pulmonary disease caused by Mycobacterium tuberculosis. Despite M. tuberculosis was identified since the evolution of medical microbiology, TB remains a global burden responsible for high rates of morbidity and mortality

[1]. The current anti-TB chemotherapy recommended by the World Health Organization

(WHO) consists of a combination of at least three first-line antibiotics, along with direct observation of patient adherence to ensure effective treatment [2]. However, premature termination of treatment due to the long duration (at least six months) and associated adverse effects contributed to an alarming rate of drug-resistant cases. The treatment of drug-resistant

TB demands second-line or third-line drugs that are often expensive, less effective, and potentially fatal [3]. Unfortunately, progress in developing new anti-TB drugs is insufficient to address the rampant emergence of multi-drug resistant (MDR) and extensively drug-resistant

(XDR) M. tuberculosis strains [4-6]. Repurposing of drugs approved for other diseases could be a promising and rapid alternate strategy to achieve new effective anti-TB drugs.

Thioridazine is a phenothiazine drug approved by the U.S. Food and Drug Administration

(FDA) for the treatment of schizophrenia and psychosis. It is also inexpensive and readily available worldwide [7]. Thioridazine has been known for several decades to possess anti- mycobacterial activity, but it is not currently being used because of its potential toxicity [8, 9].

It has multiple mechanisms of action against M. tuberculosis. First, thioridazine inhibits bacterial efflux pump to reduce phenotypic drug resistance in M. tuberculosis [10-12]. Efflux pumps are transporter proteins embedded within the cytoplasmic membrane of bacteria to extrude toxic substance, including drug molecules from intra-bacterial environment to extracellular environment. Over-expression of these pumps reduces the drug concentration in

130

the bacterial cytoplasm that opens a window for drug resistance. The suppression of these efflux pumps therefore enriches the drug concentration in the bacterial cytoplasm and hampers drug tolerance [10, 11]. Second, thioridazine inhibits potassium (K+) and calcium (Ca2+) efflux pumps of M. tuberculosis infected macrophages, thereby activating vacuolar hydrolases within the phagolysosome to degrade and kill the phagocytosed bacteria [13]. The inhibition of these efflux pumps also enhances the drug concentration within the macrophages for rapid killing of entrapped M. tuberculosis without killing the host cells [14, 15]. Third, thioridazine inhibits

Type-II NADH-menaquinone oxidoreductase, an enzyme that allows M. tuberculosis to persist in a dormant/nonreplicating stage in the oxygen- and nutrient-depleted environment of granuloma [16, 17]. The inhibition of this enzyme may be effective in the prevention of latent

TB infection. Finally, as is the case for some other anti-TB drugs, thioridazine interferes with

DNA replication and damages the bacterial envelope. These pleiotropic activities of thioridazine merit its consideration as a potential agent for the rapid treatment of MDR and

XDR-TB cases.

Various studies have demonstrated that thioridazine can significantly augment the in vitro anti- mycobacterial activity of many anti-TB drugs [18-20]. However, its efficacy has not been replicated in vivo [21, 22] or clinically owing to severe cardiotoxicity at efficacious doses [23,

24]. For example, de Knegt et al. reported that the addition of thioridazine (4, 8, or 16 mg/L) to isoniazid therapy (0.063 mg/L) significantly increased the rate of killing of M. tuberculosis

(Beijing-1585 strain) and prevented the emergence of isoniazid resistance in an in vitro time- kill assay [18]. On the other hand, when BALB/c mice infected with the same Beijing-1585 strain were treated intraperitoneally with isoniazid (25 mg/kg), rifampicin (10 mg/kg) and pyrazinamide (150 mg/kg), with thioridazine (3 to 18 mg/kg) no additional efficacy was observed compared to the isoniazid, rifampicin and pyrazinamide treatment group [18].

131

Similarly, only a modest drug synergism was observed when thioridazine (25 mg/kg) was orally administered with isoniazid (10 mg/kg) in BALB/c mice infected with M. tuberculosis

H37Rv [25]. In contrast, when thioridazine was administered at high oral doses of 32 and 70 mg/kg with isoniazid (10 mg/kg), rifampicin (10 mg/kg) and pyrazinamide (30 mg/kg) for two months, a significant reduction in the number of bacilli was observed in the lungs of the M. tuberculosis H37Rv infected BALB/c mice compared to the control and isoniazid, rifampicin and pyrazinamide treatment group [21]. Unfortunately, a thioridazine oral dose more than 40 mg/kg was not well-tolerated in guinea pigs and led to a rapid death which was hypothesised due to thioridazine associated cardiotoxicity [22]. Furthermore, adverse cardiac effects have been reported in XDR-TB patients administered thioridazine at an efficacious oral dose of 200 mg/day [24]. Other high dose-related serious side effects of thioridazine include arrhythmia, ventricular tachycardia, ventricular fibrillation and sudden cardiac death [7, 26]. Therefore, thioridazine appears to be efficacious only at high oral doses that are potentially fatal.

The requirement for high oral dose of thioridazine could be avoided by utilising the inhaled route of administration. Apart from achieving significantly higher local thioridazine concentrations in the lungs, pulmonary delivery offers additional benefits. First, the targeted delivery to the lungs reduces potential degradation of the drug in the acidic environment of the stomach and bypasses the first-pass metabolism [27, 28]. Second, pulmonary deposition maintains a high drug concentration in the lung which can reduce the frequency of dosing [29].

Third, the lung also provides a large surface area for rapid absorption and distribution of drug throughout the body to treat any disseminated diseases [27, 28]. These multiple advantages of inhalation route may reduce the dose of thioridazine required and related side-effects while allowing a rapid cure for TB.

132

The objective of the present work was to prepare and characterise a stable inhalable dry powder consisting of thioridazine and a first-line anti-TB drug, rifapentine. Rifapentine is a derivative of rifampicin that has a stronger antimicrobial profile and longer pulmonary half-life and has gained interest for pulmonary delivery [30]. Two combination powder formulations were produced and compared to determine an optimal formulation for further in vivo study: (1)

Formulation 1 consisted of amorphous thioridazine and crystalline rifapentine where each drug existed as distinctive particles, and (2) in Formulation 2 both thioridazine and rifapentine were formulated into combination of amorphous particles. The powders were produced at a mass ratio of 1:3 which was based on the current recommended oral dosages of thioridazine and rifapentine, respectively, with 20 % L-leucine as a stabiliser [24, 31]. The physiochemical properties, in vitro aerosol performance and storage stability, and in vitro microbiological activities including pulmonary cytotoxicity were compared for these two formulations.

133

2. MATERIALS AND METHODS

Materials

Thioridazine hydrochloride, L-leucine, potassium dihydrogen orthophosphate, dimethyl sulfoxide (DMSO), ascorbic acid and phosphate buffered saline (PBS) were from Sigma-

Aldrich, Castle Hill, Australia. Rifapentine was obtained from Hangzhou ICH IMP & Exp

Company Ltd., Hangzhou, China. Acetonitrile, methanol, and orthophosphoric acid were purchased from Ajax Finechem Pty Ltd., Taren Point, Australia. Middlebrook 7H9 and RPMI-

1640 media were from Bacto Laboratories, Mt. Pritchard, NSW, Australia.

Bacterial strain and cell culture

M. tuberculosis H37Ra (ATCC 25177), M. tuberculosis H37Rv (ATCC 27294), human monocytic cell line (THP-1), and adenocarcinoma human alveolar basal epithelial cell line

(A549) were purchased from American Type Culture Collection (ATCC), Manassas, VA. The mycobacterial strains were maintained in complete Middlebrook 7H9 media (enriched with 10

% (v/v) albumin, dextrose and catalase (ADC) supplement) at 37°C and stored in 20 % (v/v) glycerol at -80°C. The THP-1 and A549 cell lines were maintained in complete RPMI-1640 media (RPMI-1640 medium, 0.05 mM 2-mercaptoethanol, 10 % (v/v) fetal bovine serum) at

37°C in 5 % CO2.

134

Preparation of thioridazine and rifapentine formulations by spray drying

For Formulation 1, a crystalline rifapentine suspension was prepared as described by Chan et al. [32]. Briefly, rifapentine was dissolved in acetone:water (1:1, v/v) cosolvent at a concentration of 4 mg/mL and heated at 67°C to evaporate the acetone to form a crystalline rifapentine suspension. The aqueous suspension was then homogenised at 10 000 rpm for 2 minutes (Silverson L4RT High Shear Mixer, Silverson, Chesham, UK). Thioridazine hydrochloride (1:3 of thioridazine:rifapentine, w/w) and L-leucine (20 % w/w) were then added and solubilised in the homogenised rifapentine suspension. The combined drug mixture was spray dried using a B-290 Mini Spray Dryer (Buchi Labortechnik AG, Flawil, Switzerland) in an open-loop configuration using the following conditions: inlet temperature of 60°C, atomiser at 60 mm (approximately 740 L/h), aspirator at 100 % (40 m3/h), and feed rate of 2 mL/min.

For Formulation 2, thioridazine hydrochloride, rifapentine and L-leucine at the same ratios as

Formulation 1 were dissolved in a methanol:water (3:1, v/v) co-solvent. The solution was spray dried using similar parameters as described for Formulation 1, but in a closed-loop configuration with nitrogen as an inert gas.

Scanning electron microscopy

Particle morphology was examined using scanning electron microscopy (SEM) (Zeiss Ultra

Plus scanning microscope, Carl Zeiss, Oberkochen, Germany) at a 5 kV acceleration voltage.

Powders were deposited onto carbon tape and sputter-coated with 15 nm of gold using a K550X sputter coater (Quorum Emitech, Kent, UK).

135

Particle sizing

Equivalent volume diameter of the particles was determined using laser diffraction coupled with a Scirocco dry powder module (Mastersizer 2000, Malvern Instruments, Worcestershire,

UK). The dry powder was dispersed at an air pressure of 3.5 bar. All measurements were performed in triplicate. The refractive index (RI) was set at 1.61, based on an average RI of thioridazine (1.68), rifapentine (1.63) and L-leucine (1.46) at the mass ratio of 1:3:1.

X-ray powder diffractometry

The crystallinity of the powders was analysed using x-ray powder diffractometry (XRPD)

(D500; Siemens, Karlsruhe, Germany). The powder was spread into the cavity of the XRPD glass slide. The analysis was performed using Cu-Kα radiation at a generator voltage of 40 kV and a current of 33 mA using a 2θ scan method with a scan rate of 0.04°/sec from 5 - 40°.

Dynamic vapour sorption

A dynamic vapour sorption (DVS) instrument (DVS-1, Surface Measurement Systems,

London, UK) was used to study the moisture sorption behaviour of the powders. The powders were exposed to a dual moisture ramping cycle of 0 to 90 % with gradual 10 % increments of relative humidity (RH) at 25°C. The mass changes were recorded over time and equilibrium moisture content was determined by a dm/dt of less than 0.02 % per minute.

136

Dissolution profile

A Franz-cell dissolution system (Perm Gear Inc., Bethlehem) was used to study the dissolution profile of the powders [33]. Phosphate buffered saline (PBS) (pH 7.4) with 2 % (w/v) ascorbic acid as an antioxidant was used as a dissolution medium. Each cell reservoir was filled with 23 mL of PBS and accommodated with a magnetic stirrer to provide drug uniformity within the cell. The cells were connected to a water bath to maintain the temperature of the dissolution medium at 37°C.

Prior to dissolution, a capsule filled with 20 mg of powder was actuated using an Osmohaler®

(Plastiape, Italy) at an air flow of 100 L/min for 4 s into a next generation impactor (NGI) containing cellulose filter paper (Whatman® qualitative filter paper Grade 5) in stages 3 and 4 which have cut-off diameters of 2.2 and 1.3 µm, respectively. The filters holding the deposited powder were then placed on the top compartment of the Franz-cell reservoir. Sampling (500

µL) was performed at predetermined time points (15, 30, 45, 60, 90, 120, 180, 240, and 300 min) and replaced with an equivalent volume of fresh medium. At the end of the experiment

(5 h), the filter paper was removed and washed thoroughly with 3 mL of fresh medium.

Each formulation was tested in triplicate and the both drug masses were quantified using a validated high-performance liquid chromatography (HPLC) method adapted from previous studies [32, 34]. The mobile phase consisted of a 50 mM phosphate buffer (pH 3.0):acetonitrile at a ratio of 30:70 (%, v/v) and operated at an isocratic flow rate of 1 mL/min. The stationary phase was a µBondapakTM C18 (3.9 × 300 mm2) column (Waters, Milford, Massachusetts).

The HPLC system (Model LC-20, Shimadzu, Kyoto, Japan) was operated at a detection wavelength of 252 nm (SPD-20A UV/VIS detector, Shimadzu) and an injection volume of 20

137

µL. The total running time was 10 min with the retention time of 4.0 min for thioridazine and

5.1 min for rifapentine. The limit of quantification (LOQ) for thioridazine and rifapentine was

0.0005 mg/mL and 0.001 mg/mL, respectively. Calibration curves with R2 values of 1.0 for thioridazine and rifapentine were obtained with standard concentrations ranging from 0.001 to

0.1 mg/mL. The dissolution data were expressed as a percentage of drug mass in the buffer solution relative to the total mass recovery.

In vitro aerosol performance

In vitro aerodynamic performance of the powders was analysed using a multi-stage liquid impinger (MSLI) (British Pharmacopeia 2012, Copley, Nottingham, UK) coupled to a USP throat with a silicone mouthpiece adapter [32]. A methanol:water (3:1) co-solvent with 0.2 mg/mL ascorbic acid was used as a rinsing solvent. Each formulation was tested in triplicate.

Dispersions were performed at 45 ± 5 % RH and 20 ± 3°C. Prior to dispersion, the MSLI was filled with 20 mL rinsing solvent in stages 1 to 4 and fitted with a filter paper in stage 5. A size

3 hydroxypropyl methylcellulose capsule (Capsugel, Greenwood) was filled with 10 ± 0.5 mg of powder. The capsule was loaded and punctured in an Osmohaler® (Plastiape, Italy), and actuated for 4 s at a flow rate of 100 L/min (Flutus Air, Novi Systems Ltd., Dulverton, UK).

The dispersion components were washed with rinsing solvent as follows: 5 mL for the adaptor, and 10 mL each for the capsule, device, throat and filter. The lower cut-off diameters for stages

1 - 4 of the MSLI at 100 L/min are 10.4, 4.9, 2.4, and 1.2 µm, respectively. The drug quantification was performed using HPLC, as described above.

138

The total fine particle fraction (FPFtotal) was defined as the percentage of particles with an aerodynamic diameter less than 5 µm relative to the total recovered drug mass. The emitted fine particle fraction (FPFemitted) was defined as the percentage of particles with an aerodynamic diameter less than 5 µm in the aerosol relative to the total recovered drug mass excluding the capsule and device. Mass median aerodynamic diameter (MMAD) and geometric standard deviation (GSD) were calculated from the log-probability plot of the MSLI data.

Storage stability studies

To investigate storage stability, the powders were stored at 20 ± 3°C in loosely capped scintillation vials with the glass portion covered with aluminium foil to protect from light. Two humidity conditions were studied: 0 % and 60 % RH. For the 0 % RH storage condition, sample powders were placed in a desiccator contained silica gel desiccant beads to keep vacuum. For a 60 % RH environment, the powders were stored in a climate controlled cabinet. At intervals of 1 and 3 months, the in vitro aerosol performance of the powders was assessed using the aforementioned methods.

Resazurin assay

The resazurin assay was used to determine the minimum inhibitory concentration (MIC90) of the spray dried powders, unprocessed raw drugs in combination and their individual components against M. tuberculosis H37Ra (ATCC 25177) and M. tuberculosis H37Rv

(ATCC 27294). The assay was performed in 96-well plates in triplicate. The powders were dissolved in DMSO and serially diluted to the desired concentration in complete Middlebrook

7H9 media. A log phase M. tuberculosis culture (OD600 0.4-0.8) was added and incubated in

139

a humidified incubator at 37°C for 5 days. On day 5, a mixture of 0.02 % (w/v) resazurin solution (30 µL) and 20 % (v/v) Tween-80 (12.5 µL) was added to each well and incubated for an additional 48 h. Fluorescence was then measured using a BMG Labtech Polarstar Omega

(BMG Labtech, Ortenburg, Germany) at an excitation wavelength of 530 nm and an emission wavelength of 590 nm. After subtracting background fluorescence, the growth percentage was compared to the untreated negative control. Minimum inhibitory concentration (MIC90) was measured as the lowest drug concentration that showed 90 % growth inhibition compared to the untreated cells.

Intracellular inhibition assay

The human monocytic (THP-1) cells were grown in complete RPMI-1640 media (RPMI-1640 medium, 0.05 mM 2-mercaptoethanol, 10 % (v/v) fetal bovine serum) containing 50 ng/mL of

5 phorbol 12-myristate 13-acetate at a density of 1 × 10 cells/well for 48 h at 37°C in 5 % CO2 for the cells to adhere and differentiate. The cells were infected with M. tuberculosis H37Ra

(ATCC 25177) at a multiplicity of infection (MOI) of 5 for 4 h. The cells were then washed 3 times with PBS, replenished with fresh media and incubated overnight. Thioridazine and rifapentine alone and in combination (1:3) were prepared in RPMI-1640 media at various concentrations and then added to the plates and incubated for 48 h. At 24 and 48 h incubation, the cells were lysed and the viable bacterial number (colony forming units, CFU) was determined by plating onto Middlebrook 7H11 agar plates supplemented with 10 % (v/v) oleic acid, albumin, dextrose and catalase (OADC) enrichment, and 0.5 % (v/v) glycerol, using standard procedures as outlined in [35]. The concentrations were carefully selected, either not too high or too low, to show the bactericidal efficacy of the drugs in combination or alone. The tested concentrations were also confirmed non-toxic to THP-1 cells through cytotoxicity assay.

140

Cytotoxicity assay

The cytotoxicity profile of the spray dried powders, unprocessed raw drugs in combination and their individual components was measured using THP-1 and human alveolar basal carcinoma epithelial (A549) cell lines. A 96-well plate was seeded with the cells at a concentration of 5 ×

105 cells/well and incubated for 48 h to allow cell adherence and differentiation. The samples were dissolved in DMSO, diluted with distilled water and added into wells in two-fold dilution with concentration ranging from 0.00196 - 1.0 mg/mL. The plates were incubated for 4 days at

37°C in 5 % CO2. A 10 µL aliquot of 0.05 % (w/v) resazurin was added on day 4 and further incubated for 4 h. Fluorescence was then measured using a FLUOstar Omega Microplate

Reader (BMG Labtech, Ortenburg, Germany) with emission at 590 nm following excitation at

544 nm. The half maximal inhibitory concentration (IC50) was determined by comparing with untreated cells after subtracting the background fluorescence.

Statistics

All data are expressed as mean ± standard deviation (SD) with n=3. The t-test analysis of variance (Microsoft Office Excel 2007) was performed and the probability (p) values less than

0.05 were considered significant difference.

141

3. RESULTS

Particle morphology

Formulation 1 consisted of a mixture of raisin-like particles (thioridazine) and elongated crystalline (rifapentine), while Formulation 2 comprised of highly corrugated particles (Figure

1). The thioridazine particles in Formulation 1 and the particles in Formulation 2 were approximately 0.5 to 2.0 µm, whereas the rifapentine crystals in Formulation 1 were roughly

10  2  2 µm of length  width  height.

a b

Figure 1. Scanning electron micrographs of (a) Formulation 1 and (b) Formulation 2.

Particle sizing

Formulation 1 consisted of particles slightly larger than Formulation 2 with equivalent volume median diameter (D50) of 1.6 ± 0.03 µm and 1.0 ± 0.03 µm, respectively (Table 1). Both formulations have a narrow size distribution as indicated by a span value between 1.1 and 1.2.

However, the size data of the elongated rifapentine crystals of Formulation 1 should be taken

142

with caution as the results will depend on the orientation of the non-isometric particles in the laser diffraction measurements.

Table 1. Particle size distribution measured by laser diffractiona

D10 (µm) D50 (µm) D90 (µm) Span

Formulation 1 0.9 ± 0.03 1.6 ± 0.03 2.8 ± 0.10 1.2 ± 0.06

Formulation 2 0.6 ± 0.02 1.0 ± 0.03 1.8 ± 0.06 1.1 ± 0.04 a(mean ± SD, n=3)

X-ray diffractogram

Formulation 1 showed crystalline peaks at similar angles as previously reported for crystalline rifapentine but at lower intensities (Figure 2) [32]. Formulation 2 exhibited a “halo” pattern, indicating the powder was largely amorphous.

Formulation 2 Counts Formulation 1

5 10 15 20 25 30 35 40 2θ (Degrees)

Figure 2. X-ray diffractograms of Formulation 1 and Formulation 2. Crystalline peaks were demonstrated by Formulation 1 while Formulation 2 was largely amorphous.

143

Dynamic vapour sorption

Formulation 1 and Formulation 2 absorb a significant amount of water, up to 10.2 and 9.2 %, respectively (w/w) at 90 % RH (Figure 3). Both formulations also exhibited a similar reversible moisture sorption trend at both the first and second moisture sorptions with no moisture- induced recrystallisation. A slight sigmoid profile was observed in the desorption cycles of

Formulation 1 which appeared after 60 % RH, due to crystalline rifapentine which forms monohydrate upon exposure to elevated RH [32].

a 12 Cycle 1 Sorption Cycle 1 Desorption Cycle 2 Sorption Cycle 2 Desorption 10

4 Changein mass (%) 2

0 0 10 20 30 40 50 60 70 80 90 Relative humidity (%)

b 12 Cycle 1 Sorption Cycle 1 Desorption Cycle 2 Sorption Cycle 2 Desorption 10

4 Changein mass (%) 2

0 0 10 20 30 40 50 60 70 80 90 Relative humidity (%)

Figure 3. Dynamic vapour sorption (DVS) isotherm profiles of (a) Formulation 1 and (b)

Formulation 2 at RH ramped from 0 % to 90 % for 2-cycles.

144

Dissolution profile

The Franz-cell dissolution profiles of the formulations are displayed in Figure 4. Thioridazine in Formulation 1 underwent rapid dissolution with about 44.6 % drug released within the first

30 min and 63.6 % release within 60 min whereas, the release rate of thioridazine in

Formulation 2 was extremely slow with only about 15.6 % release over 60 min. On the other hand, rifapentine in Formulation 1 and Formulation 2 showed a cumulative release of 19.6 % and 30.1 %, respectively over 60 min followed by an extended release with a total drug release of 54.5 % and 58.2 %, respectively after 5 h.

100 Thioridazine (Formulation 1) Rifapentine (Formulation 1) 90 Thioridazine (Formulation 2) Rifapentine (Formulation 2) 80

30 Cumulativedrug release (%) 20

0 0 15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240 255 270 285 300 Time (Min)

Figure 4. Drug release profiles for Formulation 1 and Formulation 2 determined using Franz- cell dissolution method over a period of 5 h (mean ± SD, n=3).

145

In vitro aerosol performance

The FPFtotal and FPFemitted of both formulations were acceptable for pulmonary delivery, being in the range of 68.1 - 76.4 % and 73.8 - 90.5 %, respectively. However, Formulation 2 showed significantly higher FPFtotal and FPFemitted than Formulation 1 (p < 0.05) (Table 2). For example, the rifapentine in Formulation 2 had FFPtotal of 76.3 ± 2.0 % and FPFemitted of 90.0 ± 1.3 %, but the values were significantly lower (68.1 ± 3.1 % and 73.8 ± 3.4 %, respectively) in

Formulation 1. These results were also translated into a smaller MMAD of about 1.7 µm in

Formulation 2 compared to 2.3 - 2.5 µm in Formulation 1. However, the GSD of both formulations was around 1.7.

Formulation 2 showed a larger deposition of drugs in stage 3 and 4 and the filter stage while

Formulation 1 mostly deposited in stages 3 and 4 (Figure 5). Importantly, in both formulations, no significant differences were observed in the ratio of the two drugs in various stages of the

MSLI (p > 0.05), indicating both formulations achieved uniform distributions of thioridazine and rifapentine.

146

Table 2. FPFtotal, FPFemitted, MMAD, and GSD of Formulation 1 and Formulation 2 stored in 0

% RH for 3 months and dispersed using Osmohaler® at 100 L/mina

Formulation 1 0 month 1 month 3 months

Thioridazine Rifapentine Thioridazine Rifapentine Thioridazine Rifapentine

FPFtotal (%) 72.8 ± 3.2 68.1 ± 3.1 73.3 ± 3.9 67.1 ± 2.7 71.5 ± 0.7 66.4 ± 1.2

FPFemitted (%) 81.2 ± 2.1 73.8 ± 3.4 78.0 ± 4.2 71.7 ± 3.2 79.2 ± 0.8 70.0 ± 2.3

MMAD (µm) 2.3 ± 0.1 2.5 ± 0.1 2.5 ± 0.2 2.6 ± 0.1 2.5 ± 0.1 2.5 ± 0.1

GSD 1.7 ± 0.1 1.7 ± 0.1 1.7 ± 0.1 1.7 ±0.1 1.7 ± 0.1 1.7 ± 0.1

Formulation 2 0 month 1 month 3 months

Thioridazine Rifapentine Thioridazine Rifapentine Thioridazine Rifapentine

FPFtotal (%) 76.4 ± 2.1 76.3 ± 2.0 79.2 ± 3.0 77.1 ± 1.5 76.6 ± 1.6 76.6 ± 1.2

FPFemitted (%) 90.5 ± 1.4 90.0 ± 1.3 93.4 ± 1.3 92.4 ± 0.7 92.7 ± 0.7 92.2 ± 0.9

MMAD (µm) 1.7 ± 0.1 1.7 ± 0.1 1.7 ± 0.2 1.7 ± 0.1 1.7 ± 0.1 1.7 ± 0.1

GSD 1.7 ± 0.1 1.7 ± 0.1 1.8 ± 0.1 1.8 ± 0.1 1.6 ± 0.1 1.6 ± 0.1 a (mean ± SD, n=3)

147

50 a Thioridazine (0 month) Rifapentine (0 month) Thioridazine (1 month) 40 Rifapentine (1 month) Thioridazine (3 months) Rifapentine (3 months) 30

20 Recoveredmass(%)

0 Capsule Device Adapter Throat Stage 1 (10.07 Stage 2 (5.27 Stage 3 (2.40 Stage 4 (1.32 Filter µm) µm) µm) µm) (<1.32µm)

50 b Thioridazine (0 month) Rifapentine (0 month) Thioridazine (1 month) 40 Rifapentine (1 month) Thioridazine (3 months) Rifapentine (3 months) 30

20 Recoveredmass(%)

0 Capsule Device Adapter Throat Stage 1 (10.07 Stage 2 (5.27 Stage 3 (2.40 Stage 4 (1.32 Filter µm) µm) µm) µm) (<1.32µm)

Figure 5. MSLI drug deposition profiles of the powders stored in 0 % RH and 20 ± 3°C for 0 to 3 months and dispersed using Osmohaler® at 100 L/min; (a) Formulation 1 and (b)

Formulation 2 (mean ± SD, n=3).

148

Storage stability studies

Figure 6 shows the SEM morphology of the powders stored at 60 % RH and 20 ± 3°C for a month. In Formulation 1, the thioridazine and rifapentine particles were fused together and became non-dispersible larger aggregates while the particles in Formulation 2 were collapsed, fused together and caked. This indicated that the powders stored in high humidity were unstable and unsuitable for further evaluation. On the other hand, intact morphology (as in Figure 1) was observed in the powders stored at 0 % RH. Therefore, the stability dispersion was only carried out on powders that were stored in low humidity condition.

No significant differences were measured in FPFtotal, FPFemitted, MMAD and GSD (p > 0.05) when the powders were stored in vacuum for up to 3 months (Table 2). Major changes were also not observed in the MSLI deposition patterns of the powders, except after 3 months storage, Formulation 2 showed greater deposition on stage 4 instead of the filter stage (p <

0.05) (Figure 5b). There were no significant differences in the MSLI recovery of thioridazine and rifapentine in Formulation 1. In contrast, an increase in the proportion of rifapentine relative to thioridazine was observed in Formulation 2 after 3 months storage (p > 0.05) (Table

3).

149

a b

Figure 6. SEM observation of (a) Formulation 1 and (b) Formulation 2 exposed to 60 % RH for a month at 20 ± 3°C.

Table 3. Drug percentage recovery from MSLI dispersion of Formulation 1 and Formulation

Formulation 1 Formulation 2

Thioridazine Rifapentine Thioridazine Rifapentine

(%) (%) (%) (%)

0 month 30.3 ± 3.6 69.7 ± 2.7 31.5 ± 3.5 68.5 ± 2.8

1 month 29.3 ± 4.0 70.7 ± 2.6 32.6 ± 4.8 68.4 ± 3.4

3 months 30.3 ± 2.9 69.7 ± 4.2 26.3 ± 1.7 73.7 ± 4.2 a (mean ± SD, n=3)

150

Resazurin assay

Both powder formulations possessed similar MICs against both M. tuberculosis H37Ra

(0.000625 µg/mL) and M. tuberculosis H37Rv (0.005 µg/mL). The MIC90 was similar to raw thioridazine and rifapentine in a combination of 1:3 (Table 4). The MIC90 of thioridazine alone was 12.5 µg/mL and 500 µg/mL against M. tuberculosis H37Ra and M. tuberculosis H37Rv, respectively. Rifapentine alone had a MIC90 of 0.000625 µg/mL and 0.005 µg/mL against M. tuberculosis H37Ra and M. tuberculosis H37Rv, respectively, and this was similar to the combination powders. This MIC90 was increased as the proportion of thioridazine to rifapentine increased. The excipient L-leucine did not inhibit activity of M. tuberculosis H37Ra or M. tuberculosis H37Rv at the concentrations used in the formulations.

151

Table 4. Minimum inhibitory concentration (MIC90) of the combination formulations and their individual componentsa

Compound MIC90 (µg/mL)

M. tuberculosis H37Ra M. tuberculosis H37Rv

Formulation 1 0.000625 0.005

Formulation 2 0.000625 0.005

Thioridazine alone 12.5 500

Rifapentine alone 0.000625 0.005

Thioridazine:rifapentine (1:3) 0.000625 0.005

Thioridazine:rifapentine (1:1) 0.0025 0.01

Thioridazine:rifapentine (3:1) 0.025 0.01

Thioridazine:rifapentine (10:1) 0.01 > 0.02

L-leucine > 62.5 > 1000 a Mean MIC90 of three independent experiments performed in triplicate

152

Intracellular assay

To further investigate the efficacy of the formulation against M. tuberculosis-infected macrophages, we tested the bactericidal activity of solubilised rifapentine and thioridazine on

M. tuberculosis H37Ra within macrophages. Figure 7 shows the percentage of viable M. tuberculosis H37Ra after treatment with different concentrations of thioridazine and rifapentine alone and in combination of 1:3 for 24 h and 48 h.

The intracellular killing activity of thioridazine and rifapentine in combination was significantly greater than their respective single drug treatments. On day 1 the drug combination (thioridazine 0.05 µg/mL + rifapentine 0.15 µg/mL) exhibited approximately 50

% and 16 % more bactericidal effect than the thioridazine (0.05 µg/mL) and rifapentine (0.15

µg/mL) alone, respectively. A similar trend was also observed at the other two concentrations tested. Furthermore, thioridazine and rifapentine in combination exhibited a dose-dependent bactericidal activity, i.e. on day 1 the combination (thioridazine 0.5 µg/mL + rifapentine 1.5

µg/mL) resulted in 11 % increased bacterial killing than the combination at a lower drug concentration (thioridazine 0.05 µg/mL + rifapentine 0.15 µg/mL). Complete bacterial killing was observed when the concentration of the combination further was increased (thioridazine

5.0 µg/mL + rifapentine 15.0 µg/mL).

153

a 1.2E+7 ** ** 1.0E+7 **

8.0E+6 ** ** ** ** 6.0E+6

4.0E+6

2.0E+6 Mycobacterial growth (CFU/mL) growth Mycobacterial

0.0E+0 No drug THZ (0.05) RPT (0.15) THZ+RPT THZ (0.5) RPT (1.5) THZ+RPT THZ (5.0) RPT (15.0) THZ+RPT (0.05+0.15) (0.5+1.5) (5+15) Concentration (µg/mL)

b 1.2E+7 ** 1.0E+7 ** ** ** 8.0E+6 ** **

6.0E+6

4.0E+6

2.0E+6 Mycobacterial growth (CFU/mL) growth Mycobacterial

0.0E+0 No drug THZ (0.05) RPT (0.15) THZ+RPT THZ (0.5) RPT (1.5) THZ+RPT THZ (5.0) RPT (15.0) THZ+RPT (0.05+0.15) (0.5+1.5) (5+15) Concentration (µg/mL)

Figure 7. Bactericidal effect of thioridazine (THZ) and rifapentine (RPT) at 1:3 on the recovery of M. tuberculosis H37Ra from infected macrophages treated for (a) 24 h and (b) 48 h (mean

± SD, n=3). Differences between treatments were evaluated by analysis of variance (*p < 0.05,

**p < 0.01). The vertical dotted lines separate the different combination concentrations tested.

154

Cytotoxicity profile

Both Formulation 1 and Formulation 2 showed an IC50 that was two-fold lower than rifapentine alone and two-fold higher than thioridazine alone on THP-1 and A549 cell lines (Table 5). A similar IC50 value was also observed when unprocessed raw thioridazine and rifapentine were tested in a combination of 1:3. L-leucine did not inhibit either cell lines at concentrations up to

1000 µg/mL.

Table 5. Half maximal inhibitory cytotoxic concentration (IC50) of Formulation 1, Formulation

2 and their individual components on human monocytic (THP-1) and human alveolar basal carcinoma epithelial (A549) cell linesa

IC50

THP-1 A549

Formulation 1 31.25 31.25

Formulation 2 31.25 31.25

Thioridazine alone 15.63 15.63

Rifapentine alone 62.5 62.5

Thioridazine:rifapentine (1:3) 31.25 31.25

L-leucine >1000 >1000 a Mean MIC90 of three independent experiments performed in triplicate

155

4. DISCUSSION

An inhalable drug combination powder allows deposition of multiple drugs in the pulmonary region in a single inhalation [33, 36]. However, formulating drugs with different aqueous solubility poses a significant challenge. In the present study, we formulated hydrophilic thioridazine hydrochloride and hydrophobic rifapentine into two particulate systems using different production methods. In Formulation 1, the crystalline rifapentine particles were pre- formed and spray dried with solubilised thioridazine to produce amorphous thioridazine particles that were physically attached to the elongated rifapentine crystals. In our previous study, we produced a similar inhalable formulation consisted of amorphous verapamil and rifapentine crystals which showed an FPFrecovered of more than 70 % [37]. In contrast,

Formulation 2 involved a direct spray drying of thioridazine and rifapentine solubilised in methanol:water co-solvent, resulted in corrugated amorphous rifapentine-thioridazine particles. Zhou et al. used a similar approach to produce an inhalable powder of colistin sulphate (hydrophilic) and rifampicin (hydrophobic) in combination which showed a high respirable fraction of more than 90 % [38]. Furthermore, L-leucine that has intermediate water solubility was incorporated in both powders to improve powder dispersibility and stability [39,

40]. L-leucine at 20 % w/w in a spray-dried capreomycin powder has previously been safely administered by inhalation in a Phase 1 inhaled clinical study by Dharmadhikari et al. [41].

This supports our selection of 20 % w/w L-leucine in the formulations.

Formulation 2 showed superior FPFtotal and FPFemitted compared to Formulation 1 due to higher deposition in the filter stage. This also translated into a smaller MMAD for Formulation 2 (1.6

- 1.7 µm) than Formulation 1 (2.2 - 2.6 µm). The corrugated particles of Formulation 2 likely reduced the interparticle cohesive forces to improve deagglomeration of the particles [42].

156

Furthermore, corrugated particles were reported to increase shape factor (χ) to reduce the aerodynamic diameter of the particles [43]. In contrast, in a verapamil-rifapentine formulation, it was observed that the elongated rifapentine particles act as a carrier to enhance the dispersibility of amorphous verapamil particles [37]. A similar carrier effect was also reported in an inhaled formulation consisted of amorphous colistin sulphate and crystalline rifapentine

[44]. However, this does not appear to be the case in the current study as the respirable fraction of thioridazine spray dried with 20 % leucine (FPFtotal 75.2 ± 2.54) (unreported data) was similar to thioridazine of Formulation 1 (72.8 ± 3.24), indicating the carrier effect of rifapentine crystals may be drug dependent. Furthermore, the results suggest that the corrugated particle morphology of Formulation 2 has a greater role for a high respirable fraction than the particle carrier system of Formulation 1.

Another important feature of Formulation 2 was the slower in vitro dissolution rate of thioridazine than that of Formulation 1, despite the similar dissolution rate of rifapentine in both formulations. It is possible that chemical interactions between the compounds could have contributed to the difference in the dissolution observed between the formulations. However, the FTIR spectrum of Formulation 1 matched that of Formulation 2 (unreported data), suggesting the chemical characteristics of both formulations were identical. The peaks also matched with that spray dried thioridazine with leucine and rifapentine (unreported data).

Hence there are no obvious chemical variations or alterations between these two formulations which could explain the slow dissolution of thioridazine in Formulation 2. In contrast, Chan et al. showed that spray drying of a combination of isoniazid, rifampicin, and pyrazinamide in aqueous solution containing ethanol as a cosolvent resulted in the migration of the higher molecular weight (MW) and hydrophobic molecules (rifampicin) to the surface of the particles, which was reported to slow the in vitro dissolution rate of the particles [33]. Zhou et al. also

157

showed that rifampicin precipitated onto the surface of colistin sulphate which was concentrated in the internal core of the particles [38]. On the basis of these findings, in

Formulation 2, it could be hypothesised that the hydrophobic and high MW rifapentine (MW

877.0 g/mol) became supersaturated on the surface of the particles and coated the hydrophilic and low MW thioridazine (MW 407.0 g/mol). The slower dissolving surface rifapentine therefore delayed the elution of thioridazine. This feature of Formulation 2 may encourage the phagocytosis of both rifapentine and thioridazine together, which could deliver a high concentration of both drugs inside the infected alveolar macrophages. However, this may not be the case with Formulation 1 where the thioridazine and rifapentine exist as two distinct drug particles. In this formulation, it is likely that thioridazine particles dissolve rapidly in the lining of the lungs while the slower dissolving rifapentine crystals are gradually taken up by macrophages.

Both formulations required moisture-free conditions for long-term storage stability. Although these formulations showed only a 5 % moisture uptake at 25°C and 60 % RH via DVS profile, long-term storage at this condition caused formation of non-dispersible larger powder agglomerates and caked. Zhou et al. reported that coating by rifampicin can protect hygroscopic colistin sulphate from moisture-induced caking [38]. Raula et al. further demonstrated that spray drying of salbutamol sulphate with 20 % (w/w) L-leucine protected the particles from moisture-induced fusion and recrystallisation when stored at 65 % RH for 7 days [45]. Most recently, Li et al. showed that incorporation of 10 - 20 % (w/w) of L-leucine in the spray dried disodium cromoglycate offered protection against moisture-induced aerosol deterioration when the powder was stored at 60 % and 75 % RH for 24 h [46]. When thioridazine hydrochloride was spray dried alone, the powder was very hygroscopic and caked even at the ambient humidity (45 ± 5 % RH). In contrast, crystalline and amorphous rifapentine have previously

158

been reported to retain aerosol dispersibility for at least a month when stored at 60 % RH [32].

Therefore, the protection conferred by rifapentine and L-leucine may be overwhelmed in the present powder formulations by the high hygroscopicity of thioridazine. In contrast, when these powders were stored in a moisture-free condition (vacuum) for up to 3 months, they retained their physical morphology and dispersibility with no significant changes in FPFtotal, FPFemitted,

MMAD and GSD. However, the MSLI drug quantification analysis of Formulation 2 showed that the mass percentage of thioridazine and rifapentine was changed with a significant shifting of drug deposition from filter stage to stage 4 after stored in the vacuum condition for 3 months.

Further chemical analysis is necessary to study the chemical interactions between thioridazine and rifapentine of Formulation 2 which are beyond the scope of the current study. Nonetheless,

Formulation 1 with reduced chemical interaction would be preferred for further murine studies.

Both formulations showed same MIC90 as the unprocessed raw drugs against M. tuberculosis suggesting that both powder production methods did not alter the antimycobacterial activity of the drugs. However, the combination did not show any synergistic inhibitory activity against the tested M. tuberculosis strains. The inhibition of bacterial efflux pumps by thioridazine was expected to increase intra-bacterial accumulation of rifapentine and reduce the MIC90 of rifapentine. Unexpectedly, the MIC90 of rifapentine did not decrease even with the increase in thioridazine concentration. Rodrigues et al. demonstrated that the activation of bacterial efflux pumps on isoniazid-resistant M. tuberculosis strains was highly dependent on the genotype of the bacterial strain and not the drug alone [11]. Nonetheless, the efficacy of the thioridazine in combination with other anti-TB drug is well established in vivo and in clinical trials [19, 22-

25, 47].

159

Interestingly, thioridazine in combination with rifapentine enhanced the bactericidal effect against M. tuberculosis H37Ra within macrophages compared to the individual drug treatments. This result confirms that thioridazine can enhance the accumulation of rifapentine inside macrophages to increase the killing rate of the intracellular bacteria [13, 15, 26, 48].

Therefore, the combination of thioridazine and rifapentine may have a significant activity against the dormant M. tuberculosis for treatment of latent TB [19].

The cytotoxicity of Formulation 1 and Formulation 2 was assessed on THP-derived macrophages and lung epithelial cell lines. Both formulations showed a similar IC50 on both cell lines, suggesting the particle processing methods did not alter the cytotoxicity profile of the powders. The combination formulations (IC50 of 31.25 µg/mL) showed greater cytotoxicity than rifapentine alone (IC50 of 62.5 µg/mL) but were lower than thioridazine alone (15.63

µg/mL). However, it should be noted that the cytotoxicity of the formulations was tested on the solubilised form of the drugs. The cytotoxicity of the formulations therefore could be underestimated as these particles may not undergo a rapid dissolution after in vivo administration. Nonetheless, a murine pharmacokinetic study previously demonstrated that rifapentine was well-tolerated when intratracheally administered at a dose of 20 mg/kg with a subsequent pulmonary Cmax of 321 µg/mL [29].

160

5. CONCLUSION

We have demonstrated a novel inhalable powder composed of amorphous thioridazine and crystalline rifapentine (Formulation 1) with good aerosol performance and potent in vitro antimycobacterial activity. The powder required low moisture storage conditions for long-term stability. These data warrant further in vivo studies to establish the pharmacokinetic and pharmacodynamic profile of this formulation for treatment of TB.

161

ACKNOWLEDGEMENT

Thaigarajan Parumasivam is a Malaysian Government Scholarship fellow. The authors acknowledge Australian Microscopy & Microanalysis Research Facility at the Australian

Centre for Microscopy and Microanalysis, The University of Sydney for their technical support. The work was supported by the Australian Research Council’s Discovery Project

(DP120102778 and DPI150103953), the National Health and Medical Research Council

Centre of Research Excellence in Tuberculosis Control (APP1043225) and Project Grant

(APP104434), and the NSW Government Infrastructure Grant to the Centenary Institute.

162

REFERENCES

[1] World Health Organization, Global tuberculosis report 2014, in, http://www.who.int/tb/publications/global_report/gtbr14_main_text.pdf?ua=1, 2013. [2] World Health Organization, Treatment of tuberculosis: guidelines. 4th edition, 2010. [3] J.A. Caminero, G. Sotgiu, A. Zumla, G.B. Migliori, Best drug treatment for multidrug- resistant and extensively drug-resistant tuberculosis, Lancet Infect Dis, 10 (2010) 621-629. [4] L. D'Ambrosio, R. Centis, G. Sotgiu, E. Pontali, A. Spanevello, G.B. Migliori, New anti- tuberculosis drugs and regimens: 2015 update, ERJ Open Res, 1 (2015) 1-15. [5] A.I. Zumla, S.H. Gillespie, M. Hoelscher, P.P.J. Philips, S.T. Cole, I. Abubakar, T.D. McHugh, M. Schito, M. Maeurer, A.J. Nunn, New antituberculosis drugs, regimens, and adjunct therapies: needs, advances, and future prospects, Lancet Infect Dis, 14 (2014) 327-340. [6] K. Mdluli, T. Kaneko, A. Upton, The tuberculosis drug discovery and development pipeline and emerging drug targets, Cold Spring Harb Perspect Med, 5 (2015) 1-24. [7] J.G. Reilly, S.A. Ayis, I.N. Ferrier, S.J. Jones, S.H.L. Thomas, Thioridazine and sudden unexplained death in psychiatric in-patients, Br J Psychiatry, 180 (2002) 515-522. [8] H. Pleasure, Chlorpromazine (thorazine) for mental illness in the presence of pulmonary tuberculosis, Psychiatr Q, 30 (1956) 23-30. [9] Molnár J, Béládi I, F. I., Studies on antituberculotic action of some phenothiazine derivatives in vitro, Zentralbl Bakteriol Orig A, 239 (1977) 521-526. [10] D. Machado, I. Couto, J. Perdigão, L. Rodrigues, I. Portugal, P. Baptista, B. Veigas, L. Amaral, M. Viveiros, Contribution of efflux to the emergence of isoniazid and multidrug resistance in Mycobacterium tuberculosis, PLoS ONE, 7 (2012) e34538. [11] L. Rodrigues, D. Machado, I. Couto, L. Amaral, M. Viveiros, Contribution of efflux activity to isoniazid resistance in the Mycobacterium tuberculosis complex, Infect Genet Evol, 12 (2012) 695-700. [12] L. Rodrigues, D. Wagner, M. Viveiros, D. Sampaio, I. Couto, M. Vavra, W.V. Kern, L. Amaral, Thioridazine and chlorpromazine inhibition of ethidium bromide efflux in Mycobacterium avium and Mycobacterium smegmatis, J Antimicrob Chemother, 61 (2008) 1076-1082. [13] L. Amaral, M. Martins, M. Viveiros, Enhanced killing of intracellular multidrug-resistant Mycobacterium tuberculosis by compounds that affect the activity of efflux pumps, J Antimicrob Chemother, 59 (2007) 1237-1246. [14] K.K. Kishore, G.S.H. Ramakanth, N. Ch, rasekhar, P.V. Kishan, C.U. Kumar, P. Usharani, Thioridazine: a potential adjuvant in pharmacotherapy of drug resistant tuberculosis, Int J Basic Clin Pharmacol, 3 (2014) 928-932. [15] L. Amaral, J. Molnar, Mechanisms by which thioridazine in combination with antibiotics cures extensively drug-resistant infections of pulmonary tuberculosis, In Vivo, 28 (2014) 267- 271. [16] E.A. Weinstein, T. Yano, L.-S. Li, D. Avarbock, A. Avarbock, D. Helm, A.A. McColm, K. Duncan, J.T. Lonsdale, H. Rubin, Inhibitors of type II NADH:menaquinone oxidoreductase represent a class of antitubercular drugs, Proc Natl Acad Sci USA, 102 (2005) 4548-4553.

163

[17] T. Yano, L.-S. Li, E. Weinstein, J.-S. Teh, H. Rubin, Steady-state kinetics and inhibitory action of antitubercular phenothiazines on Mycobacterium tuberculosis type-II NADH- menaquinone oxidoreductase (NDH-2), J Biol Chem, 281 (2006) 11456-11463. [18] G.J. de Knegt, M.T. ten Kate, D. van Soolingen, R. Aarnoutse, M.J. Boeree, I.A.J.M. Bakker-Woudenberg, J.E.M. de Steenwinkel, Enhancement of in vitro activity of tuberculosis drugs by addition of thioridazine is not reflected by improved in vivo therapeutic efficacy, Tuberculosis, 94 (2014) 701-707. [19] A. Singh, S. Sharma, Chemotherapeutic efficacy of thioridazine as an adjunct drug in a murine model of latent tuberculosis, Tuberculosis, 94 (2014) 695-700. [20] M. Martins, Z. Schelz, A. Martins, J. Molnar, G. Hajös, Z. Riedl, M. Viveiros, I. Yalcin, E. Aki-Sener, L. Amaral, In vitro and ex vivo activity of thioridazine derivatives against Mycobacterium tuberculosis, Int J Antimicrob Agents, 29 (2007) 338-340. [21] D. van Soolingen, R. Hernandez-Pando, H. Orozco, D. Aguilar, C. Magis-Escurra, L. Amaral, J. van Ingen, M.J. Boeree, The antipsychotic thioridazine shows promising therapeutic activity in a mouse model of multidrug-resistant tuberculosis, PLoS ONE, 5 (2010) e12640. [22] N.K. Dutta, M.L. Pinn, M. Zhao, M.A. Rudek, P.C. Karakousis, Thioridazine lacks bactericidal activity in an animal model of extracellular tuberculosis, Antimicrob Agents Chemother, 68 (2013) 1327-1330. [23] F.U. Zarir, S. Tiyas, M.P. Lancelot, Safety and efficacy of thioridazine as salvage therapy in Indian patients with XDR-TB, Recent Pat Antiinfect Drug Discov, 6 (2011) 88-91. [24] E. Abbate, M. Vescovo, M. Natiello, M. Cufré, A. García, P. Gonzalez Montaner, M. Ambroggi, V. Ritacco, D. van Soolingen, Successful alternative treatment of extensively drug- resistant tuberculosis in Argentina with a combination of linezolid, moxifloxacin and thioridazine, J Antimicrob Chemother, 67 (2012) 473-477. [25] N.K. Dutta, M.L. Pinn, P.C. Karakousis, Reduced emergence of isoniazid resistance with concurrent use of thioridazine against acute murine tuberculosis, Antimicrob Agents Chemother, 58 (2014) 4048-4053. [26] L. Amaral, J. Molnar, Potential therapy of multidrug-resistant and extremely drug-resistant tuberculosis with thioridazine, In Vivo, 26 (2012) 231-236. [27] N.R. Labiris, M.B. Dolovich, Pulmonary drug delivery. Part II: The role of inhalant delivery devices and drug formulations in therapeutic effectiveness of aerosolized medications, Br J Clin Pharmacol, 56 (2003) 600-612. [28] N.R. Labiris, M.B. Dolovich, Pulmonary drug delivery. Part I: physiological factors affecting therapeutic effectiveness of aerosolized medications, Br J Clin Pharmacol, 56 (2003) 588-599. [29] J.G.Y. Chan, A.S. Tyne, A. Pang, A.J. McLachlan, V. Perera, J.C.Y. Chan, W.J. Britton, H.K. Chan, C.C. Duke, P.M. Young, D. Traini, Murine pharmacokinetics of rifapentine delivered as an inhalable dry powder, Int J Antimicrob Agents, 43 (2014) 319-323. [30] J.G.Y. Chan, X. Bai, D. Traini, An update on the use of rifapentine for tuberculosis therapy, Expert Opin Drug Deliv, 11 (2014) 421-431. [31] I.M. Rosenthal, M. Zhang, K.N. Williams, C.A. Peloquin, S. Tyagi, A.A. Vernon, W.R. Bishai, R.E. Chaisson, J.H. Grosset, E.L. Nuermberger, Daily dosing of rifapentine cures tuberculosis in three months or less in the murine model, PLoS Med, 4 (2007) e344.

164

[32] J.G.Y. Chan, C.C. Duke, H.X. Ong, J.C.Y. Chan, A.S. Tyne, H.-K. Chan, W.J. Britton, P.M. Young, D. Traini, A novel inhalable form of rifapentine, J Pharm Sci, 103 (2014) 1411- 1421. [33] J.G.Y. Chan, H.-K. Chan, C.A. Prestidge, J.A. Denman, P.M. Young, D. Traini, A novel dry powder inhalable formulation incorporating three first-line anti-tubercular antibiotics, Eur J Pharm Biopharm, 83 (2013) 285-292. [34] J. Karpinska, B. Starczewska, Simultaneous LC determination of some antidepressants combined with neuroleptics, J Pharm Biomed Anal, 29 (2002) 519-525. [35] R. Pinto, B.M. Saunders, L.R. Camacho, W.J. Britton, B. Gicquel, J.A. Triccas, Mycobacterium tuberculosis defective in phthiocerol dimycocerosate translocation provides greater protective immunity against tuberculosis than the existing bacille Calmette-Guérin vaccine, J Infect Dis, 189 (2004) 105-112. [36] J. Chan, A. Tyne, A. Pang, H.-K. Chan, P. Young, W. Britton, C. Duke, D. Traini, A rifapentine-containing inhaled triple antibiotic formulation for rapid treatment of tubercular infection, Pharm Res, 31 (2013) 1239-1253. [37] T. Parumasivam, J.G.Y. Chan, A. Pang, D.H. Quan, J.A. Triccas, W.J. Britton, H.K. Chan, In vitro evaluation of inhalable verapamil-rifapentine particles for tuberculosis therapy, Mol Pharm, 13 (2016) 979-989. [38] Q. Zhou, T. Gegenbach, J.A. Denman, H. Yu, J. Li, H.-K. Chan, Synergistic antibiotic combination powders of colistin and rifampicin provide high aerosolization efficiency and moisture protection, AAPS J, 16 (2014) 37-47. [39] T. Rattanupatam, T. Srichana, Budesonide dry powder for inhalation: effects of leucine and mannitol on the efficiency of delivery, Drug Deliv, 21 (2014) 397-405. [40] S. Adi, H. Adi, P. Tang, D. Traini, H.-k. Chan, P.M. Young, Micro-particle corrugation, adhesion and inhalation aerosol efficiency, Eur J Pharm Sci, 35 (2008) 12-18. [41] A.S. Dharmadhikari, M. Kabadi, B. Gerety, A.J. Hickey, P.B. Fourie, E. Nardell, Phase I, single-dose, dose-escalating study of inhaled dry powder capreomycin: a new approach to therapy of drug-resistant tuberculosis, Antimicrob Agents Chemother, 57 (2013) 2613-2319. [42] N. Chew, H.-K. Chan, Use of solid corrugated particles to enhance powder aerosol performance, Pharm Res, 18 (2001) 1570-1577. [43] N.Y.K. Chew, P. Tang, H.-K. Chan, J.A. Raper, How much particle surface corrugation is sufficient to improve aerosol performance of powders?, Pharm Res, 22 (2005) 148-152. [44] Q. Zhou, S.-P. Sun, J.G.Y. Chan, P. Wang, N. Barraud, S.A. Rice, J. Wang, J. Li, H.-K. Chan, Novel inhaled combination powder containing amorphous colistin and crystalline rifapentine with enhanced antimicrobial activities against planktonic cells and biofilm of Pseudomonas aeruginosa for respiratory infections, Mol Pharm, 12 (2015) 2594-2603. [45] J. Raula, F. Thielmann, J. Kansikas, S. Hietala, M. Annala, J. Seppälä, A. Lähde, E.I. Kauppinen, Investigations on the humidity-induced transformations of salbutamol sulphate particles coated with L-leucine, Pharm Res, 25 (2008) 2250-2261. [46] L. Li, S. Sun, T. Parumasivam, J.A. Denman, T. Gengenbach, P. Tang, S. Mao, H.-K. Chan, l-Leucine as an excipient against moisture on in vitro aerosolization performances of highly hygroscopic spray-dried powders, Eur J Pharm Biopharm, 102 (2016) 132-141.

165

[47] N.K. Dutta, M.L. Pinn, P.C. Karakousis, Sterilizing activity of thioridazine in combination with the first-line regimen against acute murine tuberculosis, Antimicrob Agents Chemother, 58 (2014) 5567-5569. [48] L. Amaral, J.E. Kristiansen, M. Viveiros, J. Atouguia, Activity of phenothiazines against antibiotic-resistant Mycobacterium tuberculosis: a review supporting further studies that may elucidate the potential use of thioridazine as anti-tuberculosis therapy, J Antimicrob Chemother, 47 (2001) 505-511.

166

CHAPTER 4

RIFAPENTINE-LOADED PLGA MICROPARTICLES FOR

TUBERCULOSIS INHALED THERAPY:

PREPARATION AND IN VITRO AEROSOL

CHARACTERISATION

This chapter has been published in European Journal of Pharmaceutical Sciences, 2016; 88: 1-

11, with the title ‘Rifapentine-loaded PLGA microparticles for tuberculosis inhaled therapy: preparation and in vitro aerosol characterization’. Authors: Parumasivam T, Leung SSY, Quan

DH, Triccas JA, Britton WJ, Chan H-K.

167

ABSTRACT

Inhaled delivery of drugs incorporated into poly (lactic-co-glycolic acid) (PLGA) microparticles allows a sustained lung concentration and encourages phagocytosis by alveolar macrophages that harbouring Mycobacterium tuberculosis. However, limited data are available on the effects of physicochemical properties of PLGA, including the monomer ratio

(lactide:glycolide) and molecular weight (MW) on the aerosol performance, macrophage uptake, and toxicity profile. The present study aims to address this knowledge gap, using

PLGAs with monomer ratios of 50:50, 75:25 and 85:15, MW ranged 24 - 240 kDa and an anti- tuberculosis (TB) drug, rifapentine. The PLGA-rifapentine powders were produced through a solution spray drying technique. The particles were spherical with a smooth surface and a volume median diameter around 2 µm (span ~2). When the powders were dispersed using an

® Osmohaler at 100 L/min for 2.4 s, the fine particle fraction (FPFtotal, wt. % particles in aerosol

< 5 µm relative to the total recovered drug mass) was ranged between 52 and 57 %, with no significant difference between the formulations. This result suggests that the monomer ratio and MW are not crucial parameters for the aerosol performance of PLGA. The phagocytosis analysis was performed using THP-1 monocyte-derived macrophages. The highest rate of uptake was observed in PLGA 85:15 followed by 75:25 and 50:50 with about 90 %, 80 % and

70 %, respectively phagocytosis over 4 h of exposure. Furthermore, the cytotoxicity analysis on THP-1 and human lung adenocarcinoma epithelial cells demonstrated that PLGA concentration up to 1.5 mg/ml, regardless of the monomer composition and MW, were non- toxic. In conclusion, the monomer ratio and MW are not crucial in determining the aerosol performance and cytotoxicity profile of PLGA however, the particles with high lactide composition have a superior tendency for macrophage uptake.

168

1. INTRODUCTION

Tuberculosis (TB) is most commonly manifested as pulmonary disease caused by

Mycobacterium tuberculosis, that remains a serious health problem worldwide, especially in

Asia and Africa [1]. The emergence of multidrug-resistant (MDR) and extensively drug- resistant (XDR) strains has further contributed to the spread of the disease [1]. In addition, the serious side effects of the current regimen of oral drugs, consisting of first-line anti-TB drugs isoniazid, rifampicin, ethambutol, and pyrazinamide, and the lengthy treatment course of more than six months have contributed to high rates of treatment noncompliance that lead to treatment failures and emergence of drug-resistant strains [2].

The direct delivery of therapeutic agents encapsulated in polymeric microparticles into the infection site, the lungs offers several advantages compared to the current oral regimen for treatment of TB. These advantages include delivery of high local drug concentration into the lungs, absorption of drugs into systemic circulation without undergoing hepatic first-pass metabolism, prolonged drug residence in the lungs, and allows uptake by alveolar macrophages that harbours M. tuberculosis [3-6]. Particularly, poly (lactic-co-glycolic acid) (PLGA) based inhalable dry powders have garnered considerable interest among researchers in recent years

[7-9]. PLGA is a copolymer chemically synthesised by polymerisation with different ratios of lactic acid and glycolic acid. For instance, PLGA 50:50 refers to a copolymer of 50 % lactic acid and 50 % glycolic acid composition. The PLGA polymer: (I) has been approved by Food and Drug Administration (FDA) for medical purposes due to its biocompatibility and biodegradation [10, 11], (II) permits extended release of encapsulated agents [7, 11, 12], and

(III) it is readily phagocytosed by alveolar macrophages [8, 13]. Moreover, the ability of the

169

macrophages to phagocytose these PLGA particles facilitates the trafficking/transport of the drug to other potential mycobacterial infection sites [14].

Numerous in vitro and in vivo studies have examined the efficacy of inhaled PLGA drug and vaccine systems for the treatment of TB. Suarez et al. showed guinea pigs insufflated with

PLGA rifampicin microparticles (PLGA 75:25, MW 85.2 kDa) resulted in a pulmonary drug concentration three times higher than the pure rifampicin treatment persisting 72 h post-dose

[15]. In a similar study, Yoshida et al. reported a measurable reduction of viable bacillary count in the lungs of M. tuberculosis-infected rats that were treated intratracheally with rifampicin- loaded PLGA microspheres (PLGA 75:25, MW 20 kDa) [16]. On the other hand, the recombinant M. tuberculosis antigen 85B (rAg85B) encapsulated in PLGA microspheres

(PLGA 75:25, MW 84.7 kDa) triggered a higher level of interleukin-2 production than the free soluble rAg85B in vitro T-cell assay [17]. This enhanced immune response is due to efficient uptake of vaccine microparticles by macrophages as opposed to the free antigen treatment.

Furthermore, the PLGA encapsulation was found to protect the immunogenic agents from degradation in vivo [18]. The prolonged release of antigens may also negate the need for a booster compared to free antigen administration [19].

Of the various monomer composition, PLGA 75:25 has often been chosen owing to its amorphous material nature that allows faster drug or protein release than PLGA 85:15 and more sustained release compared to PLGA 50:50 [20]. Studies have established that the degradation rate of PLGA is influenced by the monomer composition that determines the hydrophilicity of the polymer [10, 21, 22]. PLGA 50:50 degrade faster than PLGA 65:35, and

PLGA 65:35 degrade faster than PLGA 85:15, owing to the higher composition of hydrophilic glycolic acid monomer [10, 21, 22]. Similarly, PLGA with high MW requires a longer time to

170

degrade than those with low MW, due to longer polymer chains in high MW PLGA [10]. In contrast, the high MW PLGA resulted in gene expression for only three weeks, while the low

MW polymer extended the expression up to 6 months [23]. Unfortunately, limited data are available on the interplay of monomer and MW in the development of inhalable PLGA formulation for TB treatment.

To address this knowledge gap, we have compared the aerosol characteristics of various rifapentine-loaded PLGA particles with monomer ratio ranging from 50:50 to 85:15 and MW between 24 and 240 kDa. The investigated parameters include particulate morphology, particle size, thermal property, drug loading, moisture sorption, and in vitro aerosol performance.

Additionally, the effects of PLGA monomer ratio and MW on dissolution, alveolar macrophage uptake, and pulmonary cytotoxicity profiles were investigated. Rifapentine was used because it has been recognised as the most potent sterilising anti-TB agent and has a stronger bactericidal activity in dormant tubercle cells which makes it a suitable candidate for being delivered intracellularly using a polymeric carrier system [24]. Improved understanding of the influence of monomer ratio and MW on the aerosol properties of rifapentine polymeric microparticles could optimise the PLGA particles for pulmonary delivery to reduce the treatment time and side effects of the TB medication.

171

2. MATERIALS AND METHODS

Materials

The following materials were obtained from Sigma-Aldrich, Castle Hill, Australia: PLGA with monomer ratio and MW ranging from 50:50 to 85:15 and 24 to 240 kDa, respectively (Table

1), polyvinyl alcohol (MW 85 000 - 124 000), potassium dihydrogen orthophosphate, dimethyl sulfoxide (DMSO), phosphate-buffered saline tablet, and penicillin-streptomycin solution (10

000 units/mL). Rifapentine was obtained from Hangzhou ICH Imp & Exp Company Ltd.,

Hangzhou, China. Acetonitrile, methanol, tetrahydrofuran, dichloromethane, and orthophosphoric acid were from Ajax Finechem Pty Ltd., Taren Point, Australia.

PLGA microsphere preparation

Three batches of PLGA powders were made: Batch (1) spray dried with low drug loading (9

% w/w), batch (2) spray dried with high drug loading (23 % w/w), and batch (3) oil-in-water

(O/W) single emulsion solvent evaporation, as summarized in Table 1. The spray drying was performed using acetonitrile instead of dichloromethane [25]. In contrast, dichloromethane was used to produce emulsion particles due to its immiscibility nature in water.

Spray drying of drug solution

Rifapentine:PLGA at 1:10 (w/w) for Batch (1) and 3:10 (w/w) for Batch (2) were dissolved in acetonitrile. The total solid percentage was 2.2 % or 2.6 % (w/v), respectively. The spray drying was undertaken with a B-290 Mini spray-dryer coupled with a Buchi-295 inert-loop and Buchi-

172

296 dehumidifier (Buchi Laboratories, Flawil, Switzerland) using nitrogen as the drying gas.

The spray-drying conditions are inlet temperature of 60°C, atomiser of 60 mm (approximately

672 L/h), aspirator of 70 % and feed rate of 15 mL/min. The end products were kept in a desiccator containing silica gel at 20°C until used.

Oil-in-water (O/W) single emulsion solvent evaporation

Rifapentine:PLGA at 1:10 (w/w) were dissolved in dichloromethane, poured into 45 mL of 2

% PVA continuous phase and probe sonicated (Branson Sonifier 450, Branson Ultrasonics,

Danbury) at power 4 for 60 sec. The resulted emulsion was continuously stirred at constant speed using a magnetic stirrer (C-MAG MS 4, 150 × 105 × 260 mm (width × height × depth),

IKA®) at 750 rpm for 4 h to thoroughly remove the dichloromethane and hardening of the particles. The particles were then recovered by centrifugation (Beckman L90-K

Ultracentrifuge, Beckman Coulter Australia Pty Ltd, Lane Cove, Australia) at 9183 g-force for

15 min and washed three times with distilled water. Finally, the particles were reconstituted in

45 mL distilled water and dried using spray dryer. The spray dryer was operated in an open- loop mode using compressed air as the drying gas with following parameters: inlet temperature of 60°C, atomiser of 60 mm (approximately 672 L/h), aspirator of 100 % and feed rate of 2 mL/min. The collected powders were kept in a desiccator at 20°C until used.

173

Table 1. List of poly (lactide-co-glycolide) with lactide:glycolide monomer composition, molecular weight, viscosity data and production method

Formulation Lactide:Glycolidea Molecular weight Viscosity (dL/g), 0.1 %

(kDa)a (w/v)

in chloroform (25°C)a

Batch (1): Spray dried low drug loading (9 % w/w)

SP(1)50:50_24k 50:50 24-38 0.32-0.44

SP(1)50:50_38k 50:50 38-54 0.45-0.60

SP(1)50:50_54k 50:50 54-69 0.61-0.74

SP(1)75:25_76K 75:25 76-115 0.71-1.0

SP(1)85:15_50k 85:15 50-75 0.55-0.75

SP(1)85:15_190k 85:15 190-240 1.30-1.70

Batch (2): Spray dried high drug loading (27 % w/w)

SP(2)50:50_24k 50:50 24-38 0.32-0.44

SP(2)75:25_76K 75:25 76-115 0.71-1.0

SP(2)85:15_50k 85:15 50-75 0.55-0.75

Batch (3): O/W single emulsion solvent evaporation

EV50:50_24k 50:50 24-38 0.32-0.44

EV75:25_76K 75:25 76-115 0.71-1.0

EV85:15_50k 85:15 50-75 0.55-0.75 a Information obtained from Sigma-Aldrich [26].

174

Particle morphology

Scanning electron microscopy (SEM) was used to examine the morphology of the particles.

The powder was sprinkled onto a carbon sticky tape, sputter-coated with 15 nm gold (Quorum

Emitech, Kent, UK) and viewed under SEM (FESEM, Hitachi S4500) at an acceleration voltage of 5 kV.

Microsphere sizing

Equivalent volume diameter of the particle was measured by laser diffraction. The powder was dispersed via the Scirocco 2000 dry powder module (Malvern Instruments Ltd.,

Worcestershire, UK) at a dispersive air pressure of 4.0 bars. The measurement was performed in triplicate. The refractive index was set as 1.59.

Thermal analysis

Differential scanning calorimetry (DSC) (Mettler Toledo, Melbourne, Australia) was used to study the thermal property of the powders. About 5 mg of powder was mounted in an aluminium crucible, crimped with a perforated lid and heated from 25 to 300°C at 10°C/min.

Thermal data were analysed by STARe software V.9.0x (Mettler Toledo, Greifensee,

Switzerland).

175

Drug loading and drug loading efficiency

About 2 mg of powder was weighed and dissolved in 0.5 mL DMSO. One ml methanol and aqueous ascorbic acid as an antioxidant (2 mg/mL) at 75:25 (v/v), respectively was added to precipitate out the PLGA. The cloudy suspension was then centrifuged for 15 min at 13.4 k rpm (MiniSpin® Eppendorf, Hamburg, Germany) and the supernatant was analysed using

HPLC. The overall drug loading and drug loading efficiency were calculated using formula below:

The weight of drug in microspheres (mg) Drug loading = × 100 % The weight of microspheres (mg)

Actual drug loading (mg) Drug loading efficiency = × 100 % Theoretical drug loading (mg)

Dynamic vapour sorption

The moisture sorption profile of the powders was analysed using a dynamic vapour sorption

(DVS) instrument (DVS-1, Surface Measurement Systems, London, UK). Each sample was subjected to dual moisture ramping cycle of 0 to 90 % relative humidity (RH) with 10 % RH step increase, and the mass changes were recorded over time. Equilibrium moisture content at each RH was defined as dm/dt of 0.02 % or less per minute.

176

In vitro aerosolisation performance

A next generation impactor (NGI) (Copley Scientific, Nottingham, UK) coupled with a US

Pharmacopeia (USP) stainless steel throat and silicone adapter was used to determine in vitro aerosol performance of the powders. Three replicates were carried out for each formulation.

Before the dispersion, all NGI stages were sprayed with silicone grease to reduce particle bouncing. An amount of 10 ± 0.5 mg of powder was weighed into a hydroxypropyl methylcellulose (HPMC) capsule (size 3) and actuated through an Osmohaler® (Plastiape,

Italy) for 2.4 sec at an airflow of 100 L/min. The cut-off diameter for stages 1 - 7 of the NGI at 100 L/min were 6.24, 3.45, 2.18, 1.28, 0.73, 0.43 and 0.26 µm respectively.

One mL DMSO was used to wash capsule, device, adaptor, throat, and stage 1 - 8, respectively followed by addition of 1 mL methanol and ascorbic acid solvent (composition mentioned earlier) to precipitate out the PLGA. The suspensions were centrifuged for 15 min at 13.4 k rpm (MiniSpin® Eppendorf, Hamburg, Germany) and the supernatants were collected and analysed using high-performance liquid chromatography (HPLC). The total fine particle fraction (FPFtotal) (wt. % of particles with aerodynamic diameter less than 5 µm relative to the total recovered drug mass) and emitted dose (wt. % of particles emitted from capsule and device relative to the total recovered drug mass) were calculated. The total recovered mass was the amount of powder collected from inhaler, capsule, adaptor, throat and different stages of NGI.

177

In vitro dissolution studies

Approximately 1.5 mg of powder was suspended in a scintillation vial containing 15 mL phosphate-buffered saline (pH 7.4) with 2 mg/mL of ascorbic acid. The suspension was incubated at 37°C with constant agitation of 100 cycles/min using a water bath shaker (Labec,

Marrickville, Australia). At predetermined time points, 1 mL sample was withdrawn, and the microspheres were sedimented using centrifugation at 13.4 k rpm for 15 min (MiniSpin®

Eppendorf, Hamburg, Germany). The supernatant was collected for HPLC analysis while, the pellet redispersed in an equal volume of fresh media and transferred back into the scintillation vial to maintain the sink condition. All powders were tested in triplicate. Results are expressed as cumulative percentage drug release over time.

Drug distribution within particles

Prior to the macrophage uptake assay, the distribution of rifapentine within the microspheres was examined. About 2 mg of the spray dried low drug loading powder was weighed, added to

1 ml of methanol containing aqueous ascorbic acid (composition mentioned earlier) and vortexed to dissolve the drug on the particle surface. The suspension was then centrifuged for

15 min at 13.4 k rpm (MiniSpin® Eppendorf, Hamburg, Germany). The supernatant was removed and the pellet was dissolved in 0.5 ml DMSO. One ml of methanol and ascorbic acid solution was then added to precipitate out the PLGA. The suspension was centrifuged and the supernatant was collected. Both supernatants were analysed using HPLC. The drug on the external surface (first supernatant) and core (second supernatant) of the polymer matrix were expressed in percentage.

178

Macrophage uptake assay

A human monocytic cell line, THP-1 (ATCC, Manassas, VA) were grown in complete RPMI-

1640 media (RPMI-1640 medium, 0.05 mM 2-mercaptoethanol, 10 % fetal bovine serum) at

5 37°C in 5 % CO2. The cells were seeded in a 96 wells plate at a density of 10 cells/well using complete RPMI media containing 50 ng/mL phorbol 12-myristate 13-acetate and incubated for

48 h for the cells to differentiate.

Each well was treated with 200 µg of PLGA powders for 1 h, 4 h, and 8 h time intervals.

Rifapentine at a concentration of 20 µg/ml was included as a control in this study. Prior to addition to the plate, the PLGA powders were suspended in a complete RPMI media containing

1 % streptomycin-penicillin and water bath sonicated for 10 sec to obtain an even suspension.

Parallelly, the zeta potential values of the suspensions were measured at this stage using

Zetasizer (Nano ZS, Malvern Instruments, Worcestershire, UK) at 25°C. At a pre-specified time point, the cells were washed three times with phosphate-buffered saline to remove extracellular PLGA particles adhered to the surface of the cells and lysed with DMSO. The samples were collected in an Eppendorf tube, followed by precipitation with methanol and ascorbic acid solution as mentioned above. Intracellular drug concentration was expressed as the percentage of drug after normalised to the percentage of drug in the polymer core.

Cytotoxicity screening

The cytotoxicity profile was studied on THP-1 and A549, human lung adenocarcinoma epithelial cell lines (ATCC, Manassas, VA) as outlined by Pinto et al. [27]. Briefly, freshly grown cells dispensed at a density of 5 × 105 cells/well into a 96-well plate and incubated for

179

48 h for the cells adherence. The PLGA powder suspension was prepared as mentioned in the previous section and serially diluted with sterile distilled water. The diluted suspension was then transferred into the culture plates to obtain a final particle suspension concentration of

0.01 - 1.5 mg/mL. Two columns on each 96-well microtitre plate were kept for negative controls, which contained 10 μL distilled water instead of 10 µL sample. The plates were incubated for 4 days. Post-incubation, 10 µL of 0.05 % resazurin was added to each well. The plates were incubated for a further 4 h and fluorescence readings taken at 590 nm after excitation at 544 nm using an FLUOstar Omega Microplate Reader (BMG Labtech, Ortenburg,

Germany). After subtraction of background, the survival was calculated as a percentage of fluorescence compared with the untreated control.

High-performance liquid chromatography (HPLC)

Samples were analysed using an HPLC system (Shimadzu Corporation, Kyoto, Japan). The system consists of CBM-20A controller, LC-20AT pump, SPD-20A UV/VIS detector, SIL-

20A HT autosampler, and LCSolution software. The mobile phase with 53 % 50 mM phosphate buffer (pH 3), 42 % acetonitrile and 5 % tetrahydrofuran was run at an isocratic flow rate of 1 mL/min. The stationary phase was µBondapakTM C18 (3.9 × 300 mm2) column (Waters,

Milford, Massachusetts). The detection wavelength was 250 nm and the injection volume was

20 µL. A rifapentine standard curve was plotted from standard solutions with concentrations ranging from 0.5 - 100 ug/mL and an R2 value of 0.999 was obtained.

180

Statistic analysis

Data were expressed as mean ± SD with n = 3. The statistical significance of differences between groups was calculated using Student’s t-test, and p values < 0.05 were considered significant.

181

3. RESULTS AND DISCUSSION

Particle morphology

The SEM micrographs of the particles are shown in Figure 1. Both the low and high drug loading spray dried particles were spherical with a smooth surface (Figure 1a-d). The O/W single emulsion evaporation method produced a mixture of smooth and slightly corrugated particles (Figure 1e-f). It is clear that the PLGA monomer proportion and MW did not translate into any visible difference in the particulate physical morphology, but the method of powder production has an impact. Jafari-Nodoushan et al. demonstrated electrospraying of PLGA

50:50, MW 12 kDa dissolved in tetrahydrofuran solution produced irregular particles, and on increasing the MW of the polymer to 36 kDa, the particles became shell-like, wrinkled and cup-like forms [28]. This was explained by high MW polymer causing stronger chain entanglement in the solvent than the low MW polymer [29]. In another similar study, O’Hara and Hickey also obtained highly corrugated particles through spray drying dichloromethane solution of PLGA 75:25, MW 82.5 kDa with rifampicin [21]. These findings suggest that increasing the boiling point or likewise reducing the volatility of the dissolving solvent generates smooth particles. The boiling point of acetonitrile (~82°C) is relatively higher than tetrahydrofuran (~66°C) and dichloromethane (39.6°C). The slow evaporation of acetonitrile therefore allows the polymer chains at the surface of the generated droplets entangled to form a strong skin layer that overcomes the contraction pressure of further drying and hence produced smooth particles. In contrast, the weak skin layer produced by quick evaporation collapses into wrinkled or cup-like under the contraction pressure [30].

182

a b

c d

e f

Figure 1. Morphology of (a) SP(1)50:50_24k, (b) SP(1)75:25_76k, (c) SP(2)50:50_24k, (d)

SP(2)75:25_76k, (e) EV50:50_24k, and (f) EV75:25_76k.

183

Particle sizing

Table 2 shows the equivalent volume diameter analysis of the powders. The median diameter

(d50) of all particles regardless of monomer ratio, MW, and production method was around 1.3

- 1.5 µm. The d10 of SP(1)50:50_24k was slightly smaller, while d90 of SP(1)85:15_190k was slightly higher than the rest of the formulations. This result could be due to higher viscosity of

SP(1)85:15_190k which potentially caused larger droplets during atomisation of the spray drying process compared to other PLGA solutions. Likewise, SP(1)50:50_24k has a lower viscosity that resulted smaller particle. A similar observation was previously reported in chitosan based particles where a high feeding concentration with high viscosity produced larger size particles than the lower concentration solution [31, 32]. Overall, all the spray dried and

O/W particles were sized within the inhalable range (1 - 5 µm) with an unimodal distribution supporting their suitability for respiratory delivery [33].

184

Table 2. Equivalent volume diameter analysis. d10, d50 and d90 represent 10 %, 50 % and 90 % of cumulative percentage undersizea

Formulation d10 (µm) d50 (µm) d90 (µm)

SP(1)50:50_24k 0.20 ± 0.03 1.34 ± 0.05 2.64 ± 0.16

SP(1)50:50_38k 0.57 ± 0.06 1.38 ± 0.09 2.66 ± 0.12

SP(1)50:50_54k 0.67 ± 0.05 1.38 ± 0.10 2.67 ± 0.20

SP(1)75:25_76K 0.62 ± 0.05 1.34 ± 0.01 2.69 ± 0.02

SP(1)85:15_50k 0.55 ± 0.09 1.43 ± 0.11 3.21 ± 0.15

SP(1)85:15_190k 0.52 ± 0.13 1.57 ± 0.05 4.33 ± 0.31

SP(2)50:50_24k 0.68 ± 0.13 1.45 ± 0.08 2.93 ± 0.23

SP(2)75:25_76K 0.63 ± 0.02 1.60 ± 0.10 2.93 ± 0.23

SP(2)85:15_50k 0.60 ± 0.09 1.60 ± 0.08 3.37 ± 0.20

EV50:50_24k 0.58 ± 0.13 1.51 ± 0.22 3.23 ± 0.24

EV75:25_76k 0.58 ± 0.02 1.26 ± 0.03 2.54 ± 0.12

EV85:15_50k 0.53 ± 0.08 1.29 ± 0.04 2.89 ± 0.26 a mean ± SD, n = 3

185

Thermal characterisation

The glass transition temperatures (Tg) of the powders are summarised in Table 3 (see Appendix

3, Figure 1 for the thermograms). The Tg endotherm of all the PLGA powders were between

55°C and 60°C. The Tg higher than the physiological temperature (37°C) suggests the particles will have an intact morphology after deposited in the airways which makes them suitable for macrophage uptake. In addition, a decrease in Tg was observed as the composition of glycolide increased from 15 % to 50 % in all formulations. A similar trend was also observed in the unprocessed polymers. The previous study by Lan et al. showed that the incorporation of glycolide units in the crystal lattice of lactide units caused lattice defects which decreased the overall crystallinity and enthalpy of fusion of PLGA [34]. The thermograms of the present study, however, showed that the PLGA powders did not undergo melting and were most likely amorphous. Therefore, the slight decrease in the Tg may be associated with plasticiser effect caused by increasing the glycine monomer that increases the flexibility of the polymer chain

[35].

Drug loading and drug loading efficiency

A drug loading efficiency of approximately 100 % was observed for the spray dried particles, compared to less than 10 % for O/W emulsion particles (Table 3). The lower drug loading of emulsion particles was due to partitioning of the drug from dispersed oil phase into external water phase [21]. Studies have proposed various formulation strategies to overcome the limitation of low drug loading, such as using low disperse phase to continuous phase ratio and rapid solvent removal rate [36]. However, optimisation of these complex binary parameters poses a major barrier to achieve large scale production. Alternatively, high drug loading can

186

be easily achieved by spray drying method as it converts the organic solvent containing PLGA and drug directly into solid particles [37]. This single-step method is also faster, less laborious and produces high yield owing to the absence of an additional washing process to remove the surfactant, as required for the O/W method.

187

Table 3. Glass transition temperature (Tg), drug loading and drug loading efficiency of the microparticles

Formulation Glass Glass Drug loading Loading

transition transition (%) efficiency (%)

temperature temperature

(°C) (°C)

(Unprocessed

PLGA)

SP(1)50:50_24k 55.50 52.33 9.30 ± 0.6 102.0 ± 2.9

SP(1)50:50_38k 55.22 54.25 8.77 ± 0.8 99.9 ± 2.9

SP(1)50:50_54k 56.30 55.50 9.05 ± 0.8 102.5 ± 2.4

SP(1)75:25_76K 58.69 58.03 8.64 ± 0.5 95.0 ± 2.9

SP(1)85:15_50k 58.33 58.00 8.59 ± 0.1 95.0 ± 1.1

SP(1)85:15_190k 60.67 60.67 8.62 ± 0.3 94.8 ± 3.3

SP(2)50:50_24k 55.72 52.33 23.1 ± 0.2 100.1 ± 1.0

SP(2)75:25_76K 58.56 58.03 21.3 ± 0.9 99.5 ± 2.3

SP(2)85:15_50k 58.42 58.0 21.6 ± 0.3 99.2 ± 1.3

EV50:50_24k 52.29 52.33 0.72 ± 0.1 7.89 ± 1.0

EV75:25_76K 55.92 58.03 0.80 ± 0.0 8.75 ± 0.4

EV85:15_50k 57.22 58.0 0.89 ± 0.1 9.80 ± 1.3

188

Dynamic vapour sorption analysis

A slightly higher mass change was observed in SP(1)50:50_24k, while the least mass change was in both PLGA 85:15 formulations (SP(1)85:15_50k and SP(1)85:15_190k), demonstrating the hydrophobicity of the polymer increases as the lactide proportion increased (Figure 2). As shown in the literature, the lactic acid and glycolic acid share a similar chemical structure, but the presence of pendant methyl group on the alpha carbon of lactic acid makes it slightly more hydrophobic than glycolic acid [10, 38]. Therefore, increasing the monomer ratio of lactide makes the polymer less hygroscopic. Nonetheless, the minimum water uptake (< 2 %) of the powders even at a RH of 90 % suggests that the particles are stable upon exposure to high humidity. All formulations regardless of monomer ratio, drug loading, and production method also exhibited a reproducible sorption and desorption trend, demonstrating powder’s intact physicochemical stability after exposed to humidity.

189

a 1.2

0.8

0.6

0.4 SP(1)50 : 50_24k Change In Mass(%) SP(1)50 : 50_38k SP(1)50 : 50_54k 0.2 SP(1)75 : 25_76K SP(1)85 : 15_50k SP(1)85 : 15_190k 0 0 10 20 30 40 50 60 70 80 90 RH (%) b 2.2 2

1.8

1.6

1.4

1.2

0.8

Change In Mass(%) SP(2)50 : 50_24k 0.6 SP(2)75 : 25_76K SP(2)85 : 15_50k 0.4 EV50 : 50_24k 0.2 EV75 : 25_76k EV85 : 15_50k 0 0 10 20 30 40 50 60 70 80 90 RH (%)

Figure 2. First sorption cycle data from dynamic vapour sorption (DVS) isotherm of rifapentine loaded PLGA particles. The powders were exposed to 0 - 90 % RH with 10 % RH gradual increments.

190

In vitro aerosol characterisation

Table 4 compares the FPFtotal, emitted dose, MMAD and GSD of each formulation.

Representative dispersion profiles are shown in Figure 3. The monomer ratio, MW and drug loading were not crucial in determining the aerosol performance of the spray dried PLGA particles. The FPFtotal and emitted dose of low and high drug loading spray dried powders were acceptable for aerosol delivery, being in the range of 52 - 57 % and 81 - 86 %, respectively.

The MMAD and GSD were 2.5 - 3.0 µm and 2.0 - 2.5, respectively, showing the particles were within the respirable size, mildly polydisperse and suitable for deposition in the lower lung regions [39]. There were no significant differences (p > 0.05) in the FPFtotal, emitted dose,

MMAD and GSD between these formulations and this is likely due to the similar surface morphology and particle size of these formulations. Furthermore, the particles were highly charged as they readily attracted to the spatula and the storage container during the handling.

The electrostatic charge on these particles possibly outweighs the influences of the different monomer composition and MW of PLGA on the aerosol performance.

Comparing the aerosol performance of powders produced from different methods, it was noted that the O/W emulsion formulations showed a lower FPFtotal (40 - 45 %) than the corresponding solution spray dried formulations. The majority of the spray dried particles were deposited on stages 2 to 4 (Figure 3a), while particles obtained from the O/W emulsion method were primarily deposited in stage 1 (Figure 3b). The size of particles produced from these both methods was comparable. Therefore, the slightly inferior aerosol performance of the O/W emulsion particles could be attributed to the strong cohesive force between the particles.

Presumably, the turbulence inside device was unable to deagglomerate the powders into distinctive particles that resulted in bulky deposition in stage 1 of the impactor [40].

191

Coowanitwong et al. produced rifampicin PLGA (monomer ratio unspecified, MW 40 - 65 kDa) particles containing the different drug to polymer ratio via spray drying methylene chloride solution [7]. An FPF (≤ 5.8 µm) of less than 35 % regardless of drug and polymer composition was reported when the powders were dispersed by CyclohalerTM at 28.3 L/min

[7]. In another similar study, Son et al. produced PLGA (75:25; MW 66-107 kDa) coated rifampicin dihydrates crystals by spray drying a rifampicin crystalline suspension and acetone-

PLGA coating solution. The aerodynamic analysis using an Aerolizer® at 60 L/min showed the powder has an FPF (≤ 5 µm) less than 45 % [12]. In view of these findings and our data, direct solution spray drying using acetonitrile is superior in producing inhalable PLGA powder for a higher respirable dose (> 50 %) compared to spray drying chlorinated solvents or other production methods.

192

Table 4. Aerodynamic properties of various PLGA microspheres prepared using spray drying and O/W emulsion tested at a flow rate of 100 L/min for 2.4 s using an Osmohaler®,a

Formulation FPFtotal (%) Emitted dose (%) MMAD (µm) GSD

SP(1)50:50_24k 56.3 ± 3.1 85.7 ± 1.4 2.56 ± 0.1 2.03 ± 0.1

SP(1)50:50_38k 55.7 ± 3.9 81.4 ± 4.4 2.40 ± 0.4 2.02 ± 0.1

SP(1)50:50_54k 52.0 ± 1.4 83.1 ± 3.2 2.84 ± 0.2 2.01 ± 0.1

SP(1)75:25_76K 54.6 ± 1.8 81.9 ± 2.9 2.48 ± 0.4 2.03 ± 0.1

SP(1)85:15_50k 57.2 ± 1.5 85.8 ± 2.1 2.52 ± 0.4 2.05 ± 0.0

SP(1)85:15_190k 51.9 ± 1.6 85.4 ± 2.1 3.02 ± 0.2 2.65 ± 0.3

SP(2)50:50_24k 53.2 ± 4.7 80.5 ± 2.4 2.37 ± 0.1 1.95 ± 0.0

SP(2)75:25_76K 52.4 ± 3.4 82.7 ± 1.1 2.73 ± 0.1 2.01 ± 0.0

SP(2)85:15_50k 52.5 ± 1.7 82.4 ± 1.2 2.93 ± 0.1 2.10 ± 0.1

EP50:50_24k 46.8 ± 2.3 83.1 ± 0.4 2.79 ± 0.1 2.15 ± 0.1

EP75:25_76K 46.0 ± 1.6 86.3 ± 2.0 2.95 ± 0.2 2.48 ± 0.1

EP85:15_50k 40.8 ± 2.8 85.3 ± 1.5 3.08 ± 0.2 2.91 ± 0.4 a (mean ± SD, n = 3).

193

a 25 SP(1)50:50_24k

20 SP(2)50:50_24k

10 Drug recovey (%) recovey Drug

0 Capsule Device Adapter Throat Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Stage 6 Stage 7 Stage 8 (< (6.24 µm) (3.45 µm) (2.18 µm) (1.28 µm) (0.73 µm) (0.43 µm) (0.26 µm) 0.26 µm)

b 30

25 EV75:25_76k

EV85:15_50k 20

Drug recovery (%) recovery Drug 10

0 Capsule Device Adapter Throat Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Stage 6 Stage 7 Stage 8 (< (6.24 µm) (3.45 µm) (2.18 µm) (1.28 µm) (0.73 µm) (0.43 µm) (0.26 µm) 0.26 µm)

Figure 3. In vitro aerosol deposition profile following next generation impactor (NGI) dispersion at a flow rate of 100 L/min for 2.4 s using an Osmohaler® (mean ± SD, n = 3).

194

In vitro dissolution studies

The dissolution profiles of the powders are presented in Figure 4. The dissolution experiments were stopped on day 7 before a complete release (100 %) was reached because of the degradation of rifapentine in the dissolution medium after 7 days of incubation. In general, all formulations underwent a biphasic drug release; an initial burst release from the surface of the microspheres involving both dissolution and diffusion mechanisms, followed by a slow sustained release by diffusion of drug molecules from the particle core into the medium through small pores or channels in microspheres [21].

The low drug loading spray dried particles showed a burst effect in less than an hour with 23

% to 24 % drug release, while the total release was in the range of 28 % to 35 % until day 7

(Figure 4a - b). This is different from previous studies which reported more than 20 % burst release within 24 h and a total release up to 60 % within 7 days [7, 9, 21]. The much slower release observed in the present study was probably attributed to agglomeration of PLGA microspheres that reduces the particle surface area exposed to the dissolution medium (see

Appendix 3, Figure 2). Although, several surfactants including Tween 80, PVA, bovine serum albumin and polyethylene glycol at different concentrations were used to minimise the formation of PLGA agglomerates during the dissolution, the overall drug release remained similar.

The higher drug loading batch showed a slower dissolution rate than the low drug loading batch

(Figure 4b vs. Figure 4c). For instance, the maximum drug release of SP(1)85:15_50k was about 35 %, and this was reduced by more than 2-fold in SP(2)85:15_50k (13 %). This contrasts with the finding of Coowanitwong et al. [7] that degradation of high payload rifampicin-PLGA

195

particles was more rapid than the low payload particles prepared by spray drying dichloromethane-PLGA solution. A possible reason could be the decrease of the porosity or increase of the tortuosity within the particles upon spray drying with acetonitrile [7]. These phenomena possibly retarded the diffusion of dissolution media into the polymer matrix that delayed the elution of drug from the polymer. Furthermore, any differences in dissolution instrumentation setup, agitation rate, temperature or deviations from the sink conditions could have contributed to the variations in the results between studies. Currently, there is no standard in vitro dissolution method for testing inhaled formulations that can enable a reliable direct comparison between studies [41].

The burst release of O/W single emulsion solvent evaporation particles was extended until 8 h compared to less than 1 h in their counterparts in the spray dried batch (Figure 4b vs. Figure

4d). This is not surprising because the production of single emulsion particles involved washing step that removes the drugs present on the surface of the particles. However, this is the not the case with spray drying as the solvent containing the polymer and drug directly converted to dry powder which makes some portion of the drug to precipitate on the surface of the particles.

This aspect of the spray dried particles is discussed in the following section.

Despite the differences in the production method and drug loading, the dissolution of the particles followed a similar trend to that previously reported [22, 42, 43]. The polymer 85:15 had the slowest drug release followed by 75:25 and 50:50, indicating a decrease in the release kinetic with increasing lactide content. This scenario is contributed by preferential degradation of monomer glycolic acid due to higher hydrophilicity, as explained previously [10]. This observation was more profound in the higher drug loading solution spray dried and O/W emulsion formulations. In contrast, no obvious trend was noted for different MW formulations.

196

For instance, all spray dried PLGA 50:50 with different MW showed an approximately 35 % rifapentine release at the end of the experiment.

Further in vivo studies are recommended for a complete understanding on the degradation property of these polymers. The presence of enzymes and surfactants in the epithelial lung fluid perhaps may accelerate the breakdown of PLGA and also prevent agglomeration of the particles after deposition in the lungs. In addition, these particles are likely to be engulfed by alveolar macrophages, and a faster dissolution rate is expected within the macrophage environment because of the to the chelator-rich environment and acidic pH [44].

197

a 35

20 SP(1)50:50_24k SP(1)50:50_38k 15 SP(1)50:50_54k

10 SP(1)75:25_76K Cumulative Cumulative drug release (%) 5 SP(1)85:15_50k SP(1)85:15_190k 0 0 0.05 0.1 0.15 0.2 0.25 Time (Day)

b 45 40

25 SP(1)50:50_24k 20 SP(1)50:50_38k

15 SP(1)50:50_54k

10 SP(1)75:25_76K Cumulative Cumulative drug release (%) SP(1)85:15_50k 5 SP(1)85:15_190k 0 0 1 2 3 4 5 6 7 Time (Day)

198

c 25

SP(2)50:50_24k Cumulative Cumulative drug release (%) 5 SP(2)75:25_76K

SP(2)85:15_50k 0 0 1 2 3 4 5 6 7 Time (Day)

d 60

EV50:50_24k Cumulative Cumulative drug release (%) 10 EV75:25_76k EV85:15_50k 0 0 1 2 3 4 5 6 7 Time (Day)

Figure 4. Rifapentine release profiles of: (a) low drug loading spray dried PLGA powder for a duration of 8 h, (b) low drug loading spray dried PLGA powder, (c) high drug loading spray dried PLGA powder, and (d) O/W PLGA powder over 7 days (mean ± SD, n = 3).

199

Drug distribution within particles

The distribution of rifapentine in the spray dried, low drug loading PLGA microspheres is shown in Figure 5. During the spray drying process, usually the droplet shrinks inward quickly carrying any molecules in the solution. However, the study by O’Hara and Hickey [21] showed that during spray drying the PLGA molecules moved more slowly than the drug molecules. As a result, the drug precipitated throughout the dried polymer matrix, with some part of the drug available on the surface and some in the core of the particles [21]. The present study provides further insight into this mechanism. The distribution of drug was found to be a function of monomer ratio with more drugs incorporated on the surface of the particles as the glycolide content increased from 15 % to 50 %. This could be due to the variation in the hydrophobicity property of these monomers that affects the drying rate and intermolecular interaction between the drug and polymer.

100

40 External

Drugrecovery (%) 20 Internal

Figure 5. The distribution of rifapentine in spray dried PLGA particles. The mass of drug precipitated on the surface and within the core of the particles in percentage (mean ± SD, n =

3). 200

Cellular uptake and toxicity profile

Figure 6 shows the macrophage uptake of spray dried, low drug loading PLGA particles over

8 h time points. At 1 h incubation period around 50 - 60 % of particle uptake was observed regardless of polymer monomer ratio and MW. At 4 h incubation, SP(1)85:15_50k and

SP(1)85:15_190k exhibited more than 20 % drug internalisation compared to SP(1)50:50_24k,

SP(1)50:50_38k and SP(1)50:50_54k. A 100 % uptake was observed in all PLGA formulations, except SP(1)50:50_24k after 8 h incubation. The phagocytosis of particle is affected by the particle size and surface chemistry, including surface charge and hydrophobicity [45-47]. In the present study, particle size and zeta potential values (Table 5) were identical in each formulation. This confirms that the faster uptake of particles with monomer ratio 85:15 and 75:25 is due to their greater hydrophobicity that makes them have a higher tendency to adsorb to the surface of the cells [48]. In addition, for an efficient uptake, it is essential for the particles to be in nano or low micrometer range and negatively charged [45,

47]. The generated, rifapentine containing PLGA particles were in the range of 1 - 3 µm and negatively charged, and thus, they are suitable for delivering rifapentine into infected macrophages.

The lung cytotoxicity was evaluated on THP-1 and A549 cell lines using MTT assay at concentrations of 0.01 - 1.5 mg/mL for 4 days. The fact that none of the formulations led to a reduction of cell viability below 50 %, indicates that neither the different monomer ratios nor the MWs has a significant influence on the cytotoxic property of the PLGA and signifies PLGA is non-toxic to respiratory associated cells (see Appendix 3, Figure 3). These results are consistent with previous findings that PLGA is not toxic to the pulmonary cell line in vitro [8,

49].

201

120

100

60 1 h 40

4 h Drug recovery(%) 20 8 h

Figure 6. Intracellular uptake of rifapentine loaded PLGA particles by THP-1 cell line over 8 h time points (mean ± SD, n = 3). The drug recovery was expressed as percentage of total drug load after normalisation to the drug in core of the particles.

Table 5. Zeta potential of the PLGA particles dispersed in RPMI-1640 mediaa

Formulation Zeta potential (mV)

SP(1)50:50_24k -11.4 ± 0.88

SP(1)50:50_38k -11.0 ± 1.44

SP(1)50:50_54k -11.6 ± 0.71

SP(1)75:25_76K -11.7 ± 0.87

SP(1)85:15_50k -9.6 ± 1.86

SP(1)85:15_190k -10.34 ± 0.79

a (mean ± SD, n = 3).

202

4. CONCLUSION

The aerosol performance properties of the PLGA-rifapentine particles were independent of the

PLGA’s lactide:glycoside composition and MW. The highest rate of macrophage phagocytosis was observed in PLGA 85:15 followed by 75:25 and 50:50. Furthermore, the cytotoxicity analysis demonstrated that PLGA with a concentration up to 1.5 mg/ml, regardless of the monomer composition and MW were non-toxic to lung-associated cell lines, including macrophages and bronchial epithelial cells. Therefore, the PLGA-rifapentine particles provide another option for inhaled therapy for pulmonary TB.

203

ACKNOWLEDGEMENT

Thaigarajan Parumasivam is a recipient of the Malaysian Government Scholarship. Sharon

Leung is a research fellow supported by the University of Sydney. The authors acknowledge

Australian Microscopy & Microanalysis Research Facility at the Australian Centre for

Microscopy and Microanalysis, The University of Sydney for their scientific and technical assistance. The work was supported by the Australian Research Council’s Discovery Project

(DP120102778), the National Health and Medical Research Council Centre of Research

Excellence in Tuberculosis Control (APP1043225) and Project grant (APP104434), and the

NSW Government Infrastructure Grant to the Centenary Institute.

204

REFERENCES

[1] World Health Organization, Global tuberculosis report 2014, in, http://www.who.int/tb/publications/global_report/gtbr14_main_text.pdf?ua=1, 2013. [2] P. Sonnenberg, J. Murray, S. Shearer, J.R. Glynn, B. Kambashi, P. Godfrey-Faussett, Tuberculosis treatment failure and drug resistance-same strain or reinfection?, Trans R Soc Trop Med Hyg, 94 (2000) 603-607. [3] R. Pandey, G.K. Khuller, Antitubercular inhaled therapy: opportunities, progress and challenges, J Antimicrob Chemother, 55 (2005) 430-435. [4] J.P. Smith, Nanoparticle delivery of anti-tuberculosis chemotherapy as a potential mediator against drug-resistant tuberculosis, Yale J of Biol Med, 84 (2011) 361-369. [5] L. du Toit, V. Pillay, M. Danckwerts, Tuberculosis chemotherapy: current drug delivery approaches, Respir Res, 19 (2006) 118. [6] B. Semete, L. Booysen, Y. Lemmer, L. Kalombo, L. Katata, J. Verschoor, H.S. Swai, In vivo evaluation of the biodistribution and safety of PLGA nanoparticles as drug delivery systems, Nanomedicine, 6 (2010) 662-671. [7] I. Coowanitwong, V. Arya, P. Kulvanich, G. Hochhaus, Slow release formulations of inhaled rifampin, AAPS J, 10 (2008) 342-348. [8] K. Hirota, T. Hasegawa, T. Nakajima, H. Inagawa, C. Kohchi, G.-I. Soma, K. Makino, H. Terada, Delivery of rifampicin–PLGA microspheres into alveolar macrophages is promising for treatment of tuberculosis, J Control Release, 142 (2010) 339-346. [9] P. Muttil, J. Kaur, K. Kumar, A.B. Yadav, R. Sharma, A. Misra, Inhalable microparticles containing large payload of anti-tuberculosis drugs, Eur J Pharm Sci, 32 (2007) 140-150. [10] H.K. Makadia, S.J. Siegel, Poly lactic-co-glycolic acid (PLGA) as biodegradable controlled drug delivery carrier, Polymers (Basel), 3 (2011) 1377-1397. [11] A.B. Terry, A.D. Salaam, E. Nyairo, V. Thomas, D.R. Dean, PLGA nanoparticles for the sustained release of rifampicin, NGNM, 2 (2014) 1-9. [12] Y.-J. Son, J.T. McConville, Preparation of sustained release rifampicin microparticles for inhalation, J Pharm Pharmacol, 64 (2012) 1291-1302. [13] K. Ohashi, T. Kabasawa, T. Ozeki, H. Okada, One-step preparation of rifampicin/poly(lactic-co-glycolic acid) nanoparticle-containing mannitol microspheres using a four-fluid nozzle spray drier for inhalation therapy of tuberculosis, J Control Release, 135 (2009) 19-24. [14] R. Sharma, D. Saxena, A. Dwivedi, A. Misra, Inhalable microparticles containing drug combinations to target alveolar macrophages for treatment of pulmonary tuberculosis, Pharm Res, 18 (2001) 1405-1410. [15] S. Suarez, P. O'Hara, M. Kazantseva, C. Newcomer, R. Hopfer, D. McMurray, A. Hickey, Airways delivery of rifampicin microparticles for the treatment of tuberculosis, J Antimicrob Chemother, 48 (2001) 431-434. [16] A. Yoshida, M. Matumoto, H. Hshizume, Y. Oba, T. Tomishige, H. Inagawa, C. Kohchi, M. Hino, F. Ito, K. Tomoda, T. Nakajima, K. Makino, H. Terada, H. Hori, G.-I. Soma, Selective delivery of rifampicin incorporated into poly(dl-lactic-co-glycolic) acid microspheres after

205

phagocytotic uptake by alveolar macrophages, and the killing effect against intracellular Mycobacterium bovis Calmette–Guérin, Microbes Infect, 8 (2006) 2484-2491. [17] D. Lu, L. Garcia-Contreras, D. Xu, S. Kurtz, J. Liu, M. Braunstein, D. McMurray, A. Hickey, Poly (lactide-co-glycolide) microspheres in respirable sizes enhance an in vitro T cell response to recombinant Mycobacterium tuberculosis antigen 85B, Pharm Res, 24 (2007) 1834-1843. [18] D.T. O'Hagan, D. Rahman, J.P. Mcgee, H. Jeffery, M.C. Davies, P. Williams, S.S. Davis, S.J. Challacombe, Biodegradable microparticles as controlled release antigen delivery systems, Immunology, 73 (1991) 239-242. [19] K.M. Lima, S.A. Santos, V.M.F. Lima, A.A.M. Coelho-Castelo, J.M. Rodrigues Jr, C.L. Silva, Single dose of a vaccine based on DNA encoding mycobacterial hsp65 protein plus TDM-loaded PLGA microspheres protects mice against a virulent strain of Mycobacterium tuberculosis, Gene Ther, 10 (2003) 678-685. [20] D.J. Kirby, I. Rosenkrands, E.M. Agger, P. Andersen, A.G.A. Coombes, Y. Perrie, PLGA microspheres for the delivery of a novel subunit TB vaccine, J Drug Target, 16 (2008) 282- 293. [21] P. O'Hara, A. Hickey, Respirable PLGA microspheres containing rifampicin for the treatment of tuberculosis: manufacture and characterization, Pharm Res, 17 (2000) 955-961. [22] K. Makino, T. Nakajima, M. Shikamura, F. Ito, S. Ando, C. Kochi, H. Inagawa, G. Soma, H. Terada, Efficient intracellular delivery of rifampicin to alveolar macrophages using rifampicin-loaded PLGA microspheres: effects of molecular weight and composition of PLGA on release of rifampicin, Colloids Surf B Biointerfaces, 36 (2004) 35-42. [23] J.-H. Jang, L.D. Shea, Intramuscular delivery of DNA releasing microspheres: microsphere properties and transgene expression, J Control Release, 112 (2006) 120-128. [24] J.G.Y. Chan, X. Bai, D. Traini, An update on the use of rifapentine for tuberculosis therapy, Expert Opin Drug Deliv, 11 (2014) 421-431. [25] T.V.P. Doan, W. Couet, J.C. Olivier, Formulation and in vitro characterization of inhalable rifampicin-loaded PLGA microspheres for sustained lung delivery, Int J Pharm, 414 (2011) 112-117. [26] Sigma-Aldrich, Poly(lactide-co-glycolide) (PLGA) copolymers, in, Sigma-Aldrich, http://www.sigmaaldrich.com/materials-science/material-science- products.html?TablePage=112230057, 2014. [27] R. Pinto, B.M. Saunders, L.R. Camacho, W.J. Britton, B. Gicquel, J.A. Triccas, Mycobacterium tuberculosis defective in phthiocerol dimycocerosate translocation provides greater protective immunity against tuberculosis than the existing bacille Calmette-Guérin vaccine, J Infect Dis, 189 (2004) 105-112. [28] M. Jafari-Nodoushan, J. Barzin, H. Mobedi, Size and morphology controlling of PLGA microparticles produced by electro hydrodynamic atomization, Polym Adv Technol, 26 (2015) 502-513. [29] F. Wang, Z. Li, K. Tamama, C.K. Sen, J. Guan, Fabrication and characterization of prosurvival growth factor releasing, anisotropic scaffolds for enhanced mesenchymal stem cell survival/growth and orientation, Biomacromolecules, 10 (2009) 2609-2618.

206

[30] C.H. Park, J. Lee, Electrosprayed polymer particles: effect of the solvent properties, J. Appl. Polym. Sci., 114 (2009) 430-437. [31] P. He, S.S. Davis, L. Illum, Chitosan microspheres prepared by spray drying, Int J Pharm, 187 (1999) 53-65. [32] Y. Sun, F. Cui, K. Shi, J. Wang, M. Niu, R. Ma, The effect of chitosan molecular weight on the characteristics of spray-dried methotrexate-loaded chitosan microspheres for nasal administration, Drug Dev Ind Pharm, 35 (2009) 379-386. [33] W.H. Finlay, Particle size distributions, in: W.H. Finlay (Ed.) The Mechanics of Inhaled Pharmaceutical Aerosols, Academic Press, London, 2001, pp. 3-15. [34] P. Lan, Y. Zhang, Q. Gao, H. Shao, X. Hu, Studies on the synthesis and thermal properties of copoly(L-lactic acid/glycolic acid) by direct melt polycondensation, J Appl Polym Sci, 92 (2004) 2163-2168. [35] I. Edmund H, M. Herman F, Principles of plasticization, in: Plasticization and plasticizer processes, American Chemical Society, 1965, pp. 1-26. [36] N.V.N. Jyothi, P.M. Prasanna, S.N. Sakarkar, K.S. Prabha, P.S. Ramaiah, G.Y. Srawan, Microencapsulation techniques, factors influencing encapsulation efficiency, J Microencapsul, 27 (2010) 187-197. [37] J. Berggren, G. Alderborn, Effect of polymer content and molecular weight on the morphology and heat- and moisture-induced transformations of spray-dried composite particles of amorphous lactose and poly(vinylpyrrolidone), Pharm Res, 20 (2003) 1039-1046. [38] P. Gentile, V. Chiono, I. Carmagnola, P.V. Hatton, An overview of poly(lactic-co- glycolic) acid (PLGA)-based biomaterials for bone tissue engineering, Int J Mol Sci, 15 (2014) 3640-3659. [39] N.R. Labiris, M.B. Dolovich, Pulmonary drug delivery. Part I: physiological factors affecting therapeutic effectiveness of aerosolized medications, Br J Clin Pharmacol, 56 (2003) 588-599. [40] Q. Zhou, D. Morton, H. Yu, J. Jacob, J. Wang, J. Li, H.-K. Chan, Colistin powders with high aerosolisation efficiency for respiratory infection: preparation and in vitro evaluation, Immunol Rev, 102 (2013) 3736-3747. [41] T. Riley, D. Christopher, J. Arp, A. Casazza, A. Colombani, A. Cooper, M. Dey, J. Maas, J. Mitchell, M. Reiners, N. Sigari, T. Tougas, S. Lyapustina, Challenges with developing in vitro dissolution tests for orally inhaled products (OIPs), AAPS PharmSciTech, 13 (2012) 978- 989. [42] L. Lu, S.J. Peter, M. D. Lyman, H.-L. Lai, S.M. Leite, J.A. Tamada, S. Uyama, J.P. Vacanti, L. Robert, A.G. Mikos, In vitro and in vivo degradation of porous poly(DL-lactic-co- glycolic acid) foams, Biomaterials, 21 (2000) 1837-1845. [43] H. Kranz, N. Ubrich, P. Maincent, R. Bodmeier, Physicomechanical properties of biodegradable poly(D,L-lactide) and poly(D,L-lactide-co-glycolide) films in the dry and wet states, J Pharm Sci, 89 (2000) 1558-1566. [44] W.G. Kreyling, Intracellular particle dissolution in alveolar macrophages, Environ Health Perspect, 97 (1992) 121-126.

207

[45] R. Nicolete, D.F.d. Santos, L.H. Faccioli, The uptake of PLGA micro or nanoparticles by macrophages provokes distinct in vitro inflammatory response, Int Immunopharmacol, 11 (2011) 1557-1563. [46] H.S. Kruth, J. Chang, I. Ifrim, W.-Y. Zhang, Characterization of patocytosis: endocytosis into macrophage surface-connected compartments, Eur J Cell Biol, 78 (1999) 91-99. [47] S. Prior, B. Gander, N. Blarer, H.P. Merkle, M.a.L. Subirá, J.M. Irache, C. Gamazo, In vitro phagocytosis and monocyte-macrophage activation with poly(lactide) and poly(lactide- co-glycolide) microspheres, Eur J Pharm Sci, 15 (2002) 197-207. [48] R. Müller, D. Rühl, M. Lück, B.R. Paulke, Influence of fluorescent labelling of polystyrene particles on phagocytic uptake, surface hydrophobicity, and plasma protein adsorption, Pharm Res, 14 (1997) 18-24. [49] J. Wu, T. Kong, K.W.K. Yeung, H.C. Shum, K.M.C. Cheung, L. Wang, M.K.T. To, Fabrication and characterization of monodisperse PLGA-alginate core-shell microspheres with monodisperse size and homogeneous shells for controlled drug release, Acta Biomater, 9 (2013) 7410-7419.

208

CHAPTER 5

THE DELIVERY OF HIGH DOSE DRY POWDER

ANTIBIOTICS BY A LOW-COST GENERIC INHALER

This chapter has been published in The AAPS Journal, 2016; 19: 191-202, with the title ‘The delivery of high dose dry powder antibiotics by a low-cost generic inhaler’. Authors:

Parumasivam T, Leung SSY, Tang P, Mauro C, Britton WJ, Chan H-K.

209

ABSTRACT

The routine of loading multiple capsules for delivery of high dose antibiotics is time consuming which may reduce patient adherence to inhaled treatment. To overcome this limitation, an investigation was carried out using four modified versions of the Aerolizer® that accommodate a size 0 capsule for delivery of high payload formulations. In some prototypes, four piercing pins of 0.6 mm each were replaced with a single centrally located 1.2 mm pin and 1/3 reduced air inlet of the original design. The performance of these inhalers was evaluated using spray- dried antibiotic powders with distinct morphologies: spherical particles with a highly corrugated surface (colistin and tobramycin), and needle-like particles (rifapentine). The inhalers were tested at capsule loadings of 50 mg (colistin), 30 mg (rifapentine) and 100 mg

(tobramycin) using a multi-stage liquid impinger (MSLI) operating at 60 L/min. The device with a single pin and reduced air inlet showed a superior performance than the other prototypes in dispersing colistin and rifapentine powders, with a fine particle fraction (FPF wt. % < 5 µm in the aerosol) between 62 and 68 %. Subsequently, an Aerolizer® with the same configuration

(single pin and 1/3 air inlet) that accommodates a size 00 capsule was designed to increase the payload of colistin and rifapentine. The performance of the device at various inspiratory flow rates and air volumes achievable by most cystic fibrosis (CF) patients was examined at the maximum capsule loading of 100 mg. The device showed optimal performance at 45 L/min with an air volume of 1.5 – 2.0 L for colistin and 60 L/min with an air volume of 2.0 L for rifapentine. In conclusion, the modified size 00 Aerolizer® inhaler as a low-cost generic device demonstrated promising results for delivery of various high dose formulations for treatment of lung infections.

210

1. INTRODUCTION

Over the last two decades, pulmonary drug delivery has gained considerable interest as a promising alternative to conventional routes of drug administration for the treatment of various chronic pulmonary diseases, including asthma, cystic fibrosis (CF), and chronic obstructive pulmonary disease (COPD) (1-3). Nebulisers are the cornerstone of inhalation therapy which can deliver large dosages, but they are time-consuming to use, generally bulky, expensive and cause a large amount of drug wastage during continuous operation (4). The pressurised metered dose inhaler (pMDI) is compact, portable, optimal for small doses (maximum dose 5 mg), and can be used quickly (5). However, delivery efficiency of pMDI depends on the patient’s breathing pattern and hand-mouth coordination (4). The device has also suffered a phase-out of chlorofluorocarbon (CFC) and difficulties with the new hydrofluoroalkane (HFA) reformulations (6). The dry powder inhaler (DPI) is a passive device, known for its convenient use and high lung deposition, and does not require a strict hand-mouth coordination (4).

Because of these advantages, the DPI has garnered significant attention from researchers and pharmaceutical companies, and this is reflected by a plethora of dry powder-based commercial products and clinical trials (7-9).

There is a wide range of DPI products available on the market, ranging from unit-dose devices

(e.g. Foradil®/Aerolizer®, TOBI®/Podhaler® and Aridol®/Osmohaler®) to multi-dose devices

(e.g. Pulmicort®/Turbuhaler® and Advair®/Diskus®). These devices still have some drawbacks to deliver higher doses (> 100 mg) of powders, such as patients are required to load and inhale multiple capsules. TOBI®/Podhaler® marketed by Novartis for the management of

Pseudomonas aeroginosa infections in CF patients was designed to deliver a maximum dose of 112 mg tobramycin. It requires patients to load four individual capsules of 28 mg each (10).

211

Similarly, Bronchitol® by Pharmaxis for mucocilliary clearance requires patients to load and inhale up to 10 separate capsules each containing 40 mg of mannitol to generate a significant clinical outcome (11). Real-life studies from asthma and COPD patients proved that the multiple capsule refills led to poor patient adherence to inhaled medication and inferior clinical outcomes (12, 13). For instance, the Diskus® was heavily preferred over HandiHaler® by

German and Dutch patients with COPD (aged ≥ 45 years), because the former supplies a month of medication in a single drug cargo while the latter requires a capsule loading each time (13).

To overcome the hurdle of multiple capsule loading, Young et al. introduced a passive dry powder device (OrbitalTM) for delivering high dose formulations in a single dosing unit (14,

15). In vitro dispersion analysis showed the device can aerosolise mannitol co-spray dried with ciprofloxacin or azithromycin formulations at a drug loading up to 400 mg via multiple inhalation manoeuvres (14, 15). However, the device emptied a large amount of powder (nearly

100 mg) in each actuation which may cause side effects such as coughing or wheezing. Thus, there is a need for an inhaler device that can deliver large doses of medication without the requirement for multiple capsule loadings and has good tolerability in clinical settings to enhance patient compliance to inhaled treatment.

In this study, we modified the Aerolizer® to accommodate a size 0 or 00 capsule to increase the capsule drug loading. The modified Aerolizer® is a simple device consists of a mouthpiece and a chamber, therefore it is cheaper to manufacture relative to devices having complex designs or many moving parts currently available in the market. The dispersion mechanism of these modified devices is similar to the original Aerolizer® which involves air turbulence and particle impaction inside the swirling chamber for deagglomeration of the powder (16, 17).

Upon inhalation, air is drawn through the two air inlets in a tangential flow which rotates the capsule in the swirling chamber. This causes the powder to impact on the inside wall of the

212

capsule and forces the powder to empty through the capsule holes. During the exit through the holes, the powder agglomerates could be sheared into smaller fragments or individual particles and undergo further impaction with the swirling chamber wall prior to exiting the device (18).

Previous studies showed that the hole size of the capsule affects the powder emptying efficiency and dispersion. Chew et al. has reported that the capsule hole size affects the capsule and device retention, powder emptying time, respirable fraction and impaction loss of particles at the throat of the impinger using a Dinkihaler® (19). Furthermore, Coates et al. showed that the agglomeration-agglomeration interaction during the powder ejection from the capsule promotes deagglomeration of the powder using an Aerolizer® (18). The inlet size was shown to control the velocity of air entering the device which affects the level of turbulence generated inside the inhaler and intensity of the particle collision in the capsule (16, 20).

An initial investigation was carried out with four modified versions of Aerolizer® that can accommodate a size 0 capsule with the following modifications (Table 1): (1) a centrally located large piercing pin (1.2 mm) or four small pins (0.6 mm each) on each side of the capsule chamber facing the ends of the capsule, and (2) fully or 1/3 open air inlet. The dispersion performance of these devices was undertaken at a flow rate of 60 L/min and an air volume of

4.0 L using a multi-stage liquid impinger (MSLI), as suggested by Wong et al. for determining aerodynamic performance of high dose formulations (21). To simulate patient multiple breathing inspirations, the inhaler was actuated multiple times until the capsule was emptied.

Based on the obtained results, the inhaler was further optimised to accommodate a size 00 capsule to increase the capsule payload. The device was then assessed for the effect of various capsule loadings and performance at inhalation conditions considered tolerable by most CF patients. The experimental design was summarised in Figure 1.

213

Three high-dose formulations were employed to assess the performance of these prototypes.

(1) Colistin sulfate (also called polymyxin E) is a polymyxin group of antibiotics used for the treatment of Pseudomonas infection in CF patients and is often delivered through nebulisation

(22). (2) Rifapentine is a rifamycin class antibiotic given orally for the treatment of pulmonary tuberculosis (TB) and has gained significant interest to be delivered via pulmonary route to maximise the drug concentration in the lungs (23). (3) Tobramycin is an aminoglycoside antibiotic used to treat lung infections amongst CF patients via inhalation route using both nebuliser and DPI (24). The powders were produced with the spray-drying technique and characterised in terms of physiochemical properties prior to aerosol evaluation.

214

Figure 1. Flowchart summarising the experimental plan.

215

2. MATERIALS AND METHODS

Chemicals

Colistin sulfate, rifapentine and tobramycin sulfate were obtained from Hangzhou ICH IMP &

Exp Company Ltd., Hangzhou, China. L-leucine, potassium dihydrogen orthophosphate, 2,4-

Dinitrofluorobenzene, Tris(hydroxymethyl) aminomethane, dimethyl sulfoxide, and trifluoroacetic acid were from Sigma-Aldrich, Castle Hill, Australia. Acetonitrile, methanol, tetrahydrofuran, and orthophosphoric acid were purchased from Ajax Finechem Pty Ltd., Taren

Point, Australia.

Modified Aerolizer® designs and resistance measurement

The modified Aerolizer® devices were manufactured by Plastiape S.p.A. (Osnago, Italy) using methyl methacrylate-acrylonitrile-butadiene-styrene polymer for the mouthpiece, SLA® resin for the main body and mouthpiece base, and stainless steel for the internal springs and piercing needles (Figure 2). The modifications included (Table 1): (1) enlarged capsule holder to accommodate a size 0 or 00 capsule, (2) 1/3 or fully open air inlet, and (3) one large piercing pin (1.2 mm) or four small piercing pins (0.6 mm each) on each side. The original Aerolizer® can accommodate a size 3 capsule, has fully open inlet and four pins of 0.6 mm each.

The resistance of the devices was calculated using the formula (∆kPa)0.5 = QR (25). The pressure drop (kPa) across the device was measured with a pierced empty capsule at flow rates of 30, 45, 60, 80, and 100 L/min using a flow controller (Flutus Air, Novi Systems Ltd.,

Dulverton, UK) in triplicate. The square root of the pressure drop ((∆kPa)0.5 ) was then plotted

216

against corresponding flow rate (Q) and the slope of the line reported as the device resistance

(R).

Figure 2. An image of the modified Aerolizer®, 1pin_HR_0.

Table 1. Summary of Aerolizer® modifications and their resistance characteristics.

Abbreviation Capsule No. of Inlet Pin size Resistance Pressure drop

size pin size (mm) ((kPa)0.5 at 60 L/min

min/L) (kPa)

1pin_HR_0 0 1 1/3 1.2 0.022 1.8

4pin_HR_0 0 4 1/3 0.6 0.021 1.6

1pin_LR_0 0 1 Full 1.2 0.017 0.8

4pin_LR_0 0 4 Full 0.6 0.016 0.7

1pin_HR_00 00 1 1/3 1.2 0.021 1.6

Abbreviations: 1pin_HR_0, capsule size 0, 1 pin and 1/3 air inlet; 4pin_HR_0, capsule size 0,

4 pins and 1/3 air inlet; 1pin_LR_0, capsule size 0, 1 pin and full air inlet; 4pin_LR_0, capsule size 0, 4 pins and full air inlet; 1pin_HR_00, capsule size 00, 1 pin and 1/3 air inlet; HR, high resistant; LR, low resistant.

217

Dry powder preparation and physicochemical characterisation of the model inhalation particle systems

Colistin

Spray-dried colistin powder was prepared as described by Zhou et al. (26). Colistin sulfate was dissolved in distilled water at a concentration of 6 mg/mL and spray dried using B-290 mini spray dryer (Buchi Labortechnik AG, Flawil, Switzerland) in an open-loop configuration using the following settings: inlet temperature of 80°C, atomiser of 60 mm (742 L/h), aspirator of

100 % (40 m3/h), and feed rate of 2 mL/min.

Rifapentine

A crystalline rifapentine suspension was prepared as described by Chan et al. (27). Briefly, rifapentine was dissolved in acetone:water (1:1 v/v) co-solvent at a concentration of 4 mg/mL.

The solution was then heated 67°C to evaporate the acetone and form a crystalline suspension.

The suspension was then homogenised at 10 000 rpm for 2 minutes (Silverson L4RT High

Shear Mixer, Silverson, Chesham, UK). The homogenised suspension was spray dried using the aforementioned settings except the inlet temperature set at 60°C.

Tobramycin

The spray-dried tobramycin alone powder was hygroscopic and became sticky at ambient relative humidity (RH) of 45 ± 5 %. Therefore, 10 % L-leucine was incorporated as a stabiliser.

218

Tobramycin sulfate (90 % w/w) and L-leucine (10 % w/w) were dissolved in distilled water and spray dried using same conditions as the colistin powder.

Scanning electron microscopy

Particle morphology was imaged using scanning electron microscopy (SEM) (FESEM, Hitachi

S4500, Japan) at an acceleration voltage of 5 kV. Powders were deposited onto a carbon tape and sputter coated with 15 nm of gold using K550X sputter coater (Quorum Emitech, Kent,

UK).

Particle size analysis

The volume equivalent diameters (10th, 50th, and 90th percentile particle diameters) were determined using a laser diffraction coupled with a Scirocco dry powder disperser (Mastersizer

2000, Malvern Instruments, Worcestershire, UK), and the results were translated into particle size using Mie theory. The refractive index (RI) was set at 1.58, 1.63 and 1.64 for colistin, rifapentine and tobramycin, respectively, and dispersed at an air pressure of 3.5 bars (26-28).

All measurements were performed in triplicate.

X-ray powder analysis

X-ray powder diffractometer (XRD) (D500; Siemens, Karlsruhe, Germany) was used to study the crystalline property of the powders. The analysis was performed using Cu-Kα radiation at a generator voltage of 40 kV and a current of 33 mA. The data were collected by the 2θ scan method with a scan rate of 0.04°/sec from 5° to 40°.

219

Thermal analysis

Differential scanning calorimetry (DSC) (Mettler Toledo, Melbourne, Australia) was used to study the thermal profile of the dry powders. About 5 mg of powder was evenly spread into an aluminium crucible, crimped with a perforated lid and heated from 30° to 300°C at 10°C/min under a continuous flow of nitrogen gas. The perforated lid avoids explosion due to pressure build-up caused by moisture evaporation and/or sample decomposition upon heating. The thermal data was recorded and analysed through STARe software V.9.0x (Mettler Toledo,

Greifensee, Switzerland).

Dynamic vapour sorption analysis

The moisture sorption behaviour of the powders was assessed using a dynamic vapour sorption

(DVS) instrument (DVS-1, Surface Measurement Systems, London, UK). About 5 - 10 mg of powder was exposed to a dual moisture ramping cycle of 0 to 90 % with 10 % increment in RH operated at 25°C with continuous nitrogen gas flow. The mass changes were recorded over time and the equilibrium moisture content determined by dm/dt of less than 0.002 % per minute.

Bulk and tapped density determination

The powder was filled in a pre-weighed and calibrated glass tube until a volume of ~1 - 2 mL was reached. The mass of the powder was measured prior to tapping. The powder was then subjected to tapping at room condition (20 ± 5°C, 40 ± 5 % RH) until equilibrium was reached

(approximately 200 – 250 taps). The tapped volume was noted. The bulk and tapped densities

220

were calculated as the mass (g) of powder divided by the volume (cm3) of powder before and after tapping, respectively.

In vitro dispersion analysis

Aerosol performance of devices with full or 1/3 air inlet and one or four pin (s) that accommodate a size 0 capsule

The pressure drop and flow rate of 0.7 – 1.8 kPa and 60 L/min, respectively were selected as an condition achievable by most CF patients (91 – 97 % of children, 96 – 100 % of adolescents and 100 % adults) at the resistance of the modified devices (29). The multi-stage liquid impinger (MSLI) (Copley, Nottingham, UK) was operated at a flow rate of 60 L/min for 4.0 secs in a climate-controlled cabinet: temperature, 20 ± 5°C and RH, 50 ± 3 %. The capsule loading for colistin, rifapentine and tobramycin was 50 mg, 30 mg, and 100 mg, respectively with each representing an approximately 80 % capsule fill by volume of a size 0 hydroxypropyl methylcellulose (HPMC) capsule (Capsugel, NSW, Australia). The 80 % fill representing the maximum mass of the powder the capsule can accommodate. The inhaler was actuated several times until the capsule was emptied. Prior to actuation, the MSLI was calibrated using a flow controller (TSI 3063 airflow meter, TSI Instruments Ltd., Buckinghamshire, UK). Distilled water was used as a rinsing solvent for colistin and tobramycin while methanol:water at 3:1 with 2 % ascorbic acid was used for rifapentine. Each formulation was tested in triplicate and the drug mass was quantified using high-performance liquid chromatography (HPLC).

The cut-off diameters for stages 1 - 4 of the MSLI at 60 L/min were 13.0, 6.8, 3.1, and 1.7 µm, respectively. The emitted dose is the amount of drug emitted from device and capsule relative

221

to total drug recovery. The recovered fine particle fraction (FPFrecovered) was expressed as a weight percentage of drug particles with an aerodynamic diameter less than 5 µm (obtained by interpolation to the cumulative percent undersize at 5 µm) respect to the total drug recovery.

The total drug recovery was the sum of drug powder collected from the inhaler, capsule, adaptor, throat and different stages of the MSLI via HPLC quantification. The total drug recovery of all dispersions was between 95 and 105 % of the capsule loading. Mass median aerodynamic diameter (MMAD) and geometric standard deviation (GSD) were calculated from the log-probability plot of the MSLI data.

Aerosol performance of 1 pin and 1/3 air inlet open device that accommodates a size 00 capsule

(1pin_HR_00) at various capsule loadings

The capacity of 1pin_HR_00 to aerosolise colistin and rifapentine was examined at various capsule loadings of 30 mg (rifapentine) or 50 mg (colistin), 75 mg and 100 mg. The size 00

HPMC capsule reached an 80 % fill by volume at 100 mg for both powders. The 80 % fill representing the maximum mass of powder the capsule can accommodate. The dispersion conditions were kept the same as mentioned earlier.

Effect of various dispersion conditions using 1pin_HR_00

Patients are less likely to accomplish an inhaled volume of 4.0 L and this is a particular concern for CF patients who can only achieve lower inhaled volumes at the given device resistance

(29). Therefore, the dispersion efficiency of 1pin_HR_00 was tested in a range of inhaled volumes achievable by CF patient population (see Appendix 4, Table 2 and 3) at a maximum capsule loading of 100 mg colistin and rifapentine. The conditions were based on various

222

inspiratory volumes and flow rates achievable by most CF patients as reported by Tiddens et al. (29). In that study, 96 patients with varying degrees of lung disease, ages 6 - 54 years were asked to inhale forcefully through different device resistances including the resistance of

1pin_HR_00 while inspiratory flow (L/min), time (sec) and inhaled volume (L) were measured.

High-performance liquid chromatography (HPLC)

Drug quantification was performed using validated methods with an HPLC (Model LC-20,

Shimadzu, Kyoto, Japan) and UV detector (SPD-20A UV/VIS detector, Shimadzu). A linear calibration curve for colistin, rifapentine, and tobramycin was obtained using MSLI rinsing solvents with concentrations ranging from 0.005 – 2.0 mg/mL and a R2 value of 0.999.

Colistin

The method was adapted from Zhou et al. (26). The mobile phase consisted of 0.05 % trifluoroacetic acid in Mili-Q water (A) and methanol (B), and the stationary phase was a

PhenoSphere-Next 5 μm C18 (150 × 4.60 mm2) column (Phenomenex, Torrance, California).

The flow rate was 1.0 mL/min with a gradient program as follows: 30 % to 80 % (B) for 6 min, then 80 % to 30 % (B) for 3 min. The system operated at a detection of 250 nm and the injection volume was 50 µL.

Rifapentine

The method obtained from Chan et al. (27). The stationary phase was a µBondapakTM C18 (3.9

× 300 mm2) column (Waters, Milford, Massachusetts). The mobile phase was consisted of

223

phosphate buffer (pH 3.0):acetonitrile:tetrahydrofuran at 53:42:5 (% v/v). The system was operated at an isocratic flow rate of 1.0 mL/min at a wavelength of 250 nm and the injection volume was 20 µL.

Tobramycin

The method was adapted from The U.S. Pharmacopeia Convention (30). The stationary phase was Nova-Pak® C18 (3.9 × 150 mm2) column (Waters, Milford, Massachusetts). The mobile phase consisted of Tris (hydroxymethyl) aminomethane (1 .0 g), mili-Q water (400 mL), 1N sulfuric acid (10 mL) and acetonitrile (make up to 1.0 liter). An isocratic flow was set up at 1.2 mL/min at a wavelength of 365 nm and the injection volume was 20 µL.

Prior to the analysis, pre-column derivatisation was performed on the samples. An amount of

0.8 mL of samples added to 0.8 mL 2,4-Dinitrofluorobenzene (10 mg/mL in alcohol) and 2 mL

Tris(hydroxymethyl) aminomethane (3 mg/mL in dimethyl sulfoxide). The mixture heated in a water bath at 60 °C for 50 min, allowed to cool to room temperature for 10 min, followed by the addition of 2 mL acetonitrile.

Statistical analyses

All data were expressed as mean ± standard deviation (SD) with n=3. The differences were analysed using Microsoft Excel, t-test analysis of variance, and p values < 0.05 considered significant.

224

3. RESULTS

Physicochemical characterisation of the model inhalation particle systems

Scanning electron microscopy

Spray dried colistin and tobramycin particles were near-spherical with corrugated surfaces

(Figure 3a & c) (26). The rifapentine particles were elongated (Figure b), similar to that reported in the previous study (27). The geometric particle size of colistin and tobramycin ranged from 0.3 – 4.0 µm and the rifapentine crystals were roughly 10  2  2 µm of length  width  height.

Figure 3. Scanning electron microscopy observation of spray-dried (a) colistin, (b) rifapentine, and (c) tobramycin

225

Particle size analysis

The volume equivalent median diameter of the particles was within the inhalable size of less than 5 µm (see Appendix 4, Table 1). However, the size data of the elongated rifapentine should be taken with caution as the results will depend on the orientation of the non-isometric particles in the laser diffraction during measurements.

X-ray powder analysis

Both colistin and tobramycin showed a single diffuse “halo” pattern, suggesting the powders were predominantly amorphous (26, 31). In comparison, multiple crystalline peaks were exhibited by the rifapentine powder (see Appendix 4, Figure 1) (27).

Thermal analysis

Colistin and tobramycin underwent dehydration from 40°C to 170°C. Colistin showed endothermic peaks at 234.6°C and 263.7°C, responsible for melting and decomposition points, respectively. Tobramycin showed formation of a metastable anhydrous tobramycin which melted at 240.8°C, followed by melting of a stable anhydrous at 270.3°C (32). Rifapentine experienced a broad endothermic change from 30°C to 102°C and a sharp melting peak at

172.6°C (see Appendix 4, Figure 2) (27).

226

Dynamic vapour sorption analysis

Colistin and tobramycin absorbed a significant amount of water, up to 32 % and 52 %, respectively at an elevated humidity of 90 % (see Appendix 4, Figure 3). By contrast, the water absorption of crystalline rifapentine was less than 13 % even at 90 % RH. Nonetheless, all powders exhibited a similar reversible moisture sorption trend at both first and second moisture sorptions with no moisture-induced recrystallisation.

Bulk and tapped density measurement

The bulk and tapped density of the powders follow this rank: rifapentine (0.05 g/cm3 and 0.12 g/cm3, respectively) < colistin (0.08 g/cm3 and 0.17 g/cm3, respectively) < tobramycin (0.17 g/cm3 and 0.52 g/cm3, respectively).

Aerosol performance of devices with full or 1/3 air inlet and one or four pin (s) that accommodate a size 0 capsule

The results are summarised in Table 2, and the deposition profiles are depicted in Figure 4. The data are represented as an average percentage of triplicates (mean ± SD) after single or multiple actuation(s).

Colistin

All four prototype devices efficiently dispersed the colistin powder with an emitted dose of more than 70 %, and capsule and device retentions ranged between 3 - 5 % and 17 - 24 %,

227

respectively. However, the FPFrecovered of 1pin_HR_0 was significantly higher (61.8 ± 4.2 %) than the other devices (51.1 – 56.7 %; p < 0.05). As seen in Figure 4a, the 1pin_HR_0 device had a higher deposition in stage 4 and filter stage (about 50 %) compared with other devices

(< 40 %). The MMAD and GSD values of all devices ranged between 2.71 - 2.97 µm and 1.68

- 1.89, respectively.

Rifapentine

The highest emitted dose was observed in the 1pin_HR_0 (79.4 ± 1.0 %), followed by the

4pin_HR_0 (62.4 ± 2.1 %), 1pin_LR_0 (43.8 ± 3.8 %) and the lowest was in the 4pin_LR_0

(8.6 ± 0.8 %). A similar trend was also noticed in the FPFrecovered with the highest and lowest values by 1pin_HR_0 (66.1 ± 2.4 %) and 4pin_LR_0 (6.8 ± 1.4 %), respectively. The low respirable fraction of 4pin_LR_0 was caused by a substantial capsule retention of more than

80 % even after 12 actuations (Figure 4b). In contrast, 1pin_HR_0 showed drug mostly deposited in stages 3 to filter stage. The MMAD and GSD for all devices were within the range of 2.23 - 2.55 µm and 1.75 - 1.83, respectively.

Tobramycin

All devices emptied more than 80 % of the capsule loading in two actuations. However, in contrast to colistin and rifapentine powder, the highest FPFrecovered was observed in the

4pin_LR_0 device (62.7 ± 1.5 %) (p < 0.03). The calculated MMAD was comparable between all devices with the values between 2.13 and 2.23 µm while the GSD gradually decreased from

2.38 ± 0.12 (1pin_HR_0) to 1.90 ± 0.07 (4pin_LR_0). The MSLI dispersion profiles showed

228

both lower resistance devices (1pin_LR_0 and 4pin_LR_0) had about 18 % device retention compared to about 10 % in high resistance ones (1pin_HR_0 and 4pin_HR_0).

Table 2. Emitted dose, recovered fine particle fraction (FPFrecovered), mass median aerodynamic diameter (MMAD), geometric standard deviation (GSD), and number of actuations for various versions of modified Aerolizer® that accommodate a size 0 capsule tested at a flow rate 60

L/min and 4.0 L air volume.

Colistin (capsule loading 50 mg)

Emitted dose FPFrecovered MMAD (µm) GSD No. of actuations 1pin_HR_0 74.5 ± 4.3 61.8 ± 3.2* 2.84 ± 0.02 1.89 ± 0.08 1 4pin_HR_0 70.4 ± 3.8 51.8 ± 3.7 2.97 ± 0.17 1.72 ± 0.10 1 1pin_LR_0 68.1 ± 3.5 51.1 ± 3.2 2.94 ± 0.24 1.68 ± 0.04 1 4pin_LR_0 72.4 ± 3.5 56.7 ± 1.5 2.71 ± 0.12 1.69 ± 0.04 1 Rifapentine (capsule loading 30 mg)

Emitted dose FPFrecovered MMAD (µm) GSD No. of actuations 1pin_HR_0 79.4 ± 1.0 66.1 ± 2.4* 2.23 ± 0.10 1.83 ± 0.03 1 4pin_HR_0 62.4 ± 2.0 55.7 ± 4.1 2.51 ± 0.12 1.75 ± 0.10 2 1pin_LR_0 43.8 ± 3.8 37.3 ± 3.9 2.55 ± 0.10 1.78 ± 0.03 2 4pin_LR_0 8.6 ± 0.8 6.8 ± 1.4 2.55 ± 0.20 1.79 ± 0.15 12 Tobramycin (capsule loading 100 mg)

Emitted dose FPFrecovered MMAD (µm) GSD No. of actuations 1pin_HR_0 89.4 ± 1.0 56.5 ± 2.8 2.20 ± 0.09 2.38 ± 0.12 2 4pin_HR_0 90.0 ± 0.6 57.8 ± 2.4 2.13 ± 0.08 2.28 ± 0.07 2 1pin_LR_0 80.7 ± 4.7 56.1 ± 2.6 2.23 ± 0.05 2.08 ± 0.07 2 4pin_LR_0 81.1 ± 1.5 62.7 ± 1.5* 2.23 ± 0.02 1.90 ± 0.07 2 Abbreviation: 1pin_HR_0, 1 pin and 1/3 air inlet; 4pin_HR_0, 4 pins and 1/3 air inlet;

1pin_LR_0, 1 pin and full air inlet; 4pin_LR_0, 4 pins and full air inlet. (*) indicates a significant difference in FPFrecovered with p < 0.05.

229

50 a 1pin_HR_0 4pin_HR_0 40 1pin_LR_0 4pin_LR_0 30

20 Recovered mass (%) mass Recovered 10

0 Capsule Device Adaptor Throat Stage 1 (13.0 µm)Stage 2 (6.8 µm) Stage 3 (3.1 µm) Stage 4 (1.7 µm) Filter (< 1.7 µm)

1pin_HR_0 b 100 4pin_HR_0 90 1pin_LR_0 4pin_LR_0

mass (%) mass 80

Recovered Recovered 20

0 Capsule Device Adaptor Throat Stage 1 (13.0 Stage 2 (6.8 µm) Stage 3 (3.1 µm) Stage 4 (1.7 µm) Filter (< 1.7 µm) µm)

230

50 c 1pin_HR_0 4pin_HR_0 40 1pin_LR_0 4pin_LR_0 30

20 Recovered mass (%) mass Recovered 10

0 Capsule Device Adaptor Throat Stage 1 (13.0 µm)Stage 2 (6.8 µm) Stage 3 (3.1 µm) Stage 4 (1.7 µm) Filter (< 1.7 µm)

Figure 4. In vitro deposition profiles of different powders across modified Aerolizer® devices that accommodate a size 0 capsule; (a) colistin, (b) rifapentine, and (c) tobramycin at a flow rate 60 L/min and 4.0 L air volume. Abbreviations: 1pin_HR_0, 1 pin and 1/3 air inlet; 4pin_HR_0, 4 pins and 1/3 air inlet; 1pin_LR_0, 1 pin and full air inlet; 4pin_LR_0, 4 pins and full air inlet. The error bars represent the standard deviation of three measurements.

231

Aerosol performance of 1 pin and 1/3 air inlet open device that accommodates a size 00 capsule (1pin_HR_00) at various capsule loadings

In light of the high respirable fraction produced by the 1pin_HR_0 configuration dispersing colistin and rifapentine, a similar device that can accommodate a size 00 capsule was designed to increase the drug payload. The performance of the device to aerosolise colistin and rifapentine powders at capsule loadings of 30 mg to 100 mg is summarised in Table 3 with their respective dispersion profiles shown in Figure 5.

Colistin

Irrespective of drug loading, the device emitted about 75 % of powder and taking 2 to 3 actuations to empty the capsule. In contrast, FPFrecovered was significantly decreased from 61.1

± 2.6 % to 51.7 ± 2.8 % and 50.0 ± 1.3 % when the capsule loading increased from 50 mg to

75 mg and 100 mg, respectively. This trend was also reflected in the MMAD, 2.53 ± 0.07 µm for 50 mg loading while 3.04 - 3.05 µm for 75 mg and 100 mg loadings. The GSD was not significantly changed between different capsule loadings with values between 1.74 and 1.76 (p

> 0.50). Regardless of capsule loading, most particles were deposited in stage 3 to filter stage of the impinger. Approximately 21 - 25 % of drug was retained in the capsule and device.

Rifapentine

As the loading increased, the emitted dose was slightly reduced from 82.0 - 87.8 % (30 mg and

75 mg loading) to 74.3 - 80.2 % (100 mg loading) with no significant difference in FPFrecovered

(55 - 60 %) and MMAD (2.35 – 2.61 µm). The number of actuations required to empty the

232

capsule was increased from 2, 5 to 7 times as the drug loading increased from 30 mg, 75 mg to

100 mg, respectively, showing a direct relationship between dose emptying and the number of actuations. The MSLI aerosolisation profile showed the particles mostly deposited in stages 3 to filter stage with about 20 % capsule and device retention in all loadings.

233

Table 3. Emitted dose, recovered fine particle fraction (FPFrecovered), mass median aerodynamic diameter (MMAD), geometric standard deviation

(GSD), no. of actuations and various capsule loadings for the 1 pin and 1/3 air inlet device that accommodates a size 00 capsule (1pin_HR_00) tested at a flow rate 60 L/min and 4.0 L air volume.

Loading Colistin Rifapentine

……..(mg) 50 mg 75 mg 100 mg 30 mg 75 mg 100 mg

Performance

Emitted dose (%) 78.0 ± 4.3 74.0 ± 3.0 76.0 ± 0.81 84.9 ± 2.9 83.6 ± 1.0 77.2 ± 3.9

FPFrecovered (%) 61.1 ± 2.6 51.7 ± 2.8 50.0 ± 1.3 58.9 ± 2.7 60.4 ± 3.0 54.6 ± 2.6

MMAD (µm) 2.53 ± 0.07 3.05 ± 0.14 3.04 ± 0.05 2.35 ± 0.13 2.36 ± 0.22 2.61 ± 0.15

GSD 1.74 ± 0.09 1.76 ± 0.04 1.75 ± 0.04 1.82 ± 0.07 2.11 ± 0.13 1.85 ± 0.04

No. of actuations 2 2 3 2 5 7

234

50 a 50mg 75mg 40 100mg

20 Recovered mass (%) mass Recovered 10

0 Capsule Device Adaptor Throat Stage 1 (13.0 µm)Stage 2 (6.8 µm) Stage 3 (3.1 µm) Stage 4 (1.7 µm) Filter (< 1.7 µm)

b 50 30mg 75mg 40 100mg

20 Recovered mass (%) mass Recovered 10

0 Capsule Device Adaptor Throat Stage 1 (13.0 µm)Stage 2 (6.8 µm) Stage 3 (3.1 µm) Stage 4 (1.7 µm) Filter (< 1.7 µm)

Figure 5. Aerosol deposition profiles across 1 pin and 1/3 inlet device that accommodates a size 00 capsule (1pin_HR_00); (a) colistin and (b) rifapentine following MSLI at a flow rate 60 L/min and 4.0 L air volume. The error bars represent the standard deviation of three measurements.

235

Effect of various dispersion conditions using 1pin_HR_00

The optimal flow rates and air volumes tolerable by CF patients at a resistance equivalent to

1pin_HR_00 (0.021 kPa0.5 min/L) with a capsule loading of 100 mg colistin or rifapentine are presented in Table 4. Their respective dispersion profiles are presented in Figure 6. Please refer to Appendix 4 (Table 2 and 3) for the results from other dispersion conditions.

Colistin

Optimal inhalation conditions for colistin were 45 L/min for 2.0 sec and 2.7 sec that are equivalent to air volumes of 1.5 L and 2.0 L, respectively. Under these conditions, there were no significant differences in emitted dose (2.0 sec, 80.6 ± 1.4 %; 2.7 sec, 81.0 ± 3.7 %),

FPFrecovered (2.0 sec, 46.9 ± 3.3 %; 2.7 sec, 51.0 ± 1.5 %), MMAD (2.0 sec, 3.67 ± 0.09 µm; 2.7 sec, 3.45 ± 0.17 µm), and GSD (2.0 sec, 2.05 ± 0.02; 2.7 sec, 1.89 ± 0.03 %) (p > 0.05). In addition, the device had withdrawn an average of 15.3 ± 5.0 mg and 21.7 ± 10.4 mg of powder per puff, respectively and required 6 - 8 actuations to empty the capsule. The MSLI dispersion profiles showed the powder largely deposited in stages 3 and 4 of the MSLI.

Rifapentine

As for the rifapentine, optimal performance was observed at 60 L/min for 2.0 sec (inhaling 2.0

L air) with emitted dose of 83.3 ± 3.7 %, FPFrecovered of 58.4 ± 2.4 %, MMAD of 2.50 ± 0.13

µm and GSD of 2.01 ± 0.07. About 8 - 9 actuations were required to empty the capsule with an average of 9.2 ± 5.2 mg powder dispersed through the device in each actuation.

236

Table 4. Optimal inhalation conditions for efficient performance of 1 pin and 1/3 air inlet device that accommodates a size 00 capsule (1pin_HR_00) at a capsule loading of 100 mg colistin or rifapentine.

Flow rate/volume Colistin Rifapentine

45 L/min, 45 L/min, 60 L/min,

Performance 1.5 L 2.0 L 2.0 L

Emitted dose (%) 80.6 ± 1.4 81.0 ± 3.7 83.3 ± 3.7

FPFrecovered (%) 46.9 ± 3.3 51.0 ± 1.5 58.4 ± 2.4

MMAD (µm) 3.67 ± 0.09 3.45 ± 0.17 2.50 ± 0.13

GSD 2.05 ± 0.02 1.89 ± 0.03 2.01 ± 0.07

No. of actuations 8 6 8 - 9

Powder actuated in the first actuation 32.0 ± 6.9 17.0 ± 8.9 10.7 ± 3.7

(mg)

Average of powder dispersed in each 15.3 ± 5.0 21.7 ± 10.4 9.2 ± 5.2 actuation (mg)

237

40 Colistin 45 L/min, 1.5 L Colistin 45 L/min, 2.0 L Rifapentine 60 L/min, 2.0 L 30

20 Recovered mass Recovered (%)mass 10

0 Capsule Device Adaptor Throat Stage 1 Stage 2 Stage 3 Stage 4 Filter

Figure 6. Deposition profiles using the 1 pin and 1/3 air inlet device that accommodates a size

00 capsule (1pin_HR_00) with a drug loading of 100 mg colistin or rifapentine at their optimal inhalation conditions as specified in the legend. The error bars represent the standard deviation of three measurements.

238

4. DISCUSSION

The intrinsic resistance of a DPI, which affects the turbulence across the device, is an essential parameter determining how easily patients can generate certain inspiratory flow rates for optimal lung deposition (33, 34). The prototype modified Aerolizer® devices are ranked by their resistance as 1pin_HR_0 (0.022 kPa0.5 min/L) > 4pin_HR_0 and 1pin_HR_00 (0.021 kPa0.5 min/L) > 1pin_LR_0 (0.017 kPa0.5 min/L) > 4pin_LR_0 (0.016 kPa0.5 min/L). The resistance of these devices is within the range of commercially available DPIs (29, 34, 35). For instance, the original Aerolizer®, which is characterised by a low resistance of 0.017 kPa0.5 min/L, was well tolerated and able to administer a dose of 108 µg/day formoterol to patients with mild-to-moderate persistent asthma (36, 37). In contrast, RS01® HRes, a device with a high resistance of more than 0.036 kPa0.5 min/L was shown successfully operated by majority of CF patients to deliver 40 mg mannitol (11). Therefore, patients should be able to generate sufficient flow rates to operate these modified Aerolizer® devices for optimal lung deposition.

The dispersion performance of the modified devices was tested using powders that are exhibiting diverse physicochemical characteristics. Both spray-dried colistin and tobramycin particles were amorphous with corrugated surfaces and rifapentine were elongated crystals.

Colistin and rifapentine were pure drug formulations, whereas tobramycin contains 10 % L- leucine as an excipient. These formulations allowed the devices to be tested with dry powders of various chemical and physical properties.

The preliminary investigations showed that the capsule size 0 device with 1/3 air inlet and single pin configuration (1pin_HR_0) produced higher FPFrecovered (62 - 66 %) for the colistin and rifapentine powders. The comparison of the devices at the same flow rate (60 L/min)

239

resulted in markedly different pressure drops among these devices due to the different resistance - 1/3 air inlet high resistance devices produced larger pressure drop (1.6 - 1.8 kPa) than the fully open inlet low resistance devices (0.7 - 0.8 kPa). These differences in pressure drop are a strong component that drives powder fluidisation and dispersion. This is correlated with the computational fluid dynamics (CFD) analysis of the original Aerolizer® with 1/3 air inlet which was shown to generate greater turbulence, integral scale strain rate (ISSR) and velocity than the full air inlet counterpart. This combined effects therefore increased the capsule impaction inside the swirling chamber that enhanced the powder deagglomeration and generated higher respirable fraction (16, 20). On the other hand, the single piercing pin of 1.2 mm diameter was found to be more efficient in discharging powder from capsule than the four pins of 0.6 mm diameter each, although the total aperture area of both designs was same (0.36 mm2). A prominent example is the 4pin_LR_0 (fully open air inlet and 4 pins) retained nearly

80 % of the rifapentine in the capsule even after 12 actuations which was attributed to the elongated shape and larger surface area of the particles (38). In contrast, when the colistin and

® rifapentine powders were dispersed with the original Aerolizer at 100 L/min, FPFrecovered of about 83 % has been reported in both powders (26, 27). The original Aerolizer® due to its smaller size could generate a stronger flow field at the given dispersion condition which may have caused stronger powder deagglomeration and higher respirable fraction (26, 27).

Furthermore, the modified devices showed higher capsule and device retention (19 - 24 % colistin and 18 - 23 % rifapentine) compared to the original Aerolizer® (~ 10 % colistin and ~

15 % rifapentine) due to the larger size of these components.

In contrast, when tobramycin powder was dispersed using the 1pin_HR_0 device, the

FPFrecovered was significantly reduced to 56.5 ± 2.8 % (p < 0.05). The MSLI dispersion profile showed higher drug deposition in the throat and stage 1 compared to the other devices. This

240

observation suggests an incomplete deagglomeration of the powder, and this could be due to a larger fraction of the powder exiting the device before a maximum level of turbulence and impaction was generated in the swirling chamber (19, 20). Studies have reported that during spray drying, L-leucine is coated onto the surface of the drug particles to improve the flowability of the powder (39, 40). This may aid the premature emptying of the tobramycin powder prior to the maximum flow and impaction development. Note that the bulk density of tobramycin (0.17 g/cm3) powder is also higher than that of the colistin (0.08 g/cm3) and rifapentine (0.05 g/cm3), and this potentially has contributed to less powder fluidisation resulting in higher deposition of tobramycin in the throat and stage 1 of the MSLI (41). The tobramycin powder was therefore not used for further optimisation with the device having a similar configuration as the 1pin_HR_0 but can accommodate a size 00 capsule (1pin_HR_00).

Using the 1pin_HR_00 device for the dispersion of colistin and rifapentine showed that the number of actuations was significantly increased as the capsule loading increased from 30 mg to 100 mg. When the drug loading was 100 mg a slower capsule spinning was noticed, suggesting the turbulence at the given flow rate was not strong enough to efficiently spin the heavier capsule. The slow spinning of the capsule could have reduced the powder-capsule impaction and the deagglomeration of large aggregates/clumps that resulted in less powder coming out from the capsule. Similarly, at the low loading dose the powder has more room in the rotating capsule for impaction which breaks up the large agglomerates of powder to easily emit from capsule (18). This observation is consistent with the CFD analysis of the original

Aerolizer® shown that the higher particle-capsule impaction enhanced the deagglomeration potential in the flow field and improved the dispersion (18, 42). Furthermore, this impaction has caused a significant increase of FPFrecovered in colistin (50 mg vs. 75 mg and 100 mg, p <

0.01), but only a slight increase in rifapentine (30 mg vs. 75 mg and 100 mg, p > 0.05). This

241

could be due to the amorphous and more cohesive characteristics of colistin compared to rifapentine, which is crystalline and less cohesive (14, 38). These results also suggest that the performance of the device depends on powder properties (i.e. morphology and crystallinity), and therefore if different powders to colistin and rifapentine are tested the outcome could be different. Nonetheless, the FPFrecovered of both powders were still within the acceptable respirable fraction of 50 - 61 %, implying that irrespective of capsule loading the device is capable of aerosolising these powders.

When colistin was dispersed using 1pin_HR_00 at patient-tolerable inhalation conditions, an optimal performance was observed at a flow rate of 45 L/min and an air volume of 1.5 - 2.0 L.

Tiddens et al. reported that nearly all CF patients can achieve 45 L/min at a resistance equivalent to that of the 1pin_HR_00. Furthermore, about 51 - 77 % of adults (> 18 years), 29

- 62 % of adolescents (11 - 18 years), and 3 - 30 % children (< 11 years) were reported to be able to generate an inspiratory volume of 1.5 - 2.0 L (29). At these dispersion conditions, about

17 - 32 mg powder exited the device in the first actuation, the capsule was emptied in 6 - 8 actuations, and the FPFrecovered was in an acceptable range of 46.9 – 51.0 %. In contrast, the dispersion at same flow rate with a volume of 1.0 L produced a lower FPFrecovered (41.8 ± 1.4

%), indicating the inspiratory volume must be more than 1.5 L for a high respirable fraction.

Alternatively, when the flow rate was increased to 60 L/min with an air volume of 1.0 L, a larger amount of powder (~62 mg out of 100 mg colistin) was emptied in the first actuation

(see Appendix 4, Table 2), which may not be tolerable by patients. Therefore, when inhaling through the 1pin_HR_00, the flow rate should not be more than 45 L/min to avoid withdrawing an excessive amount of colistin powder. A conventional tool to identify and maintain a correct air flow includes a trainer inhaler device, such as the Diskus® trainer, which will make a whistling noise as a warning to prevent patients from inhaling over 40 L/min (43). Although

242

the trainer devices can provide an objective assessment of inhalation profile (44), it cannot be guaranteed that patients will maintain a similar flow rate when inhaling through the actual device. To overcome the drawback of high flow rate, a flow-modifying setup with 1pin_HR_00 and a flow restrictor was proposed. The device exhibited a pressure drop 1.6 kPa at a flow rate of 60 L/min without restrictor and 45 L/min with the restrictor, suggesting the spike of flow rate more than 45 L/min could be prevented by attaching the device to a restrictor. Additional dispersion analysis with this setup showed the performance of the 1pin_HR_00 (FPFrecovered,

43.3 - 46.9 %) was not significantly affected by the flow restrictor. Thus, this simple flow restriction setup opens a window to optimise the performance of 1pin_HR_00 suitable for patients.

For rifapentine, an optimal flow rate and volume were 60 L/min and 2.0 L, respectively. The percentage of CF patients who can achieve this condition is 51 % of adults, 29 % of adolescents, and 3 % of children (29). The FPFrecovered was in an acceptable range of 58.4 ± 2.4 %, an average of 10.7 ± 3.7 mg powder exited the device in the first actuation, and required 8 - 9 actuations to empty the capsule. In contrast, the dispersion analysis at 60 L/min with a volume of 1.5 L reduced the FPFrecovered to 16.0 ± 1.7 % due to high capsule retention (nearly 34 %) even after

14 actuations. Similarly, a low FPFrecovered of 14.3 ± 1.0 % was observed at 45 L/min and 2.0 L with nearly 70 % of drug retained in the capsule after 14 actuations (refer to supplementary material Table 3). Reducing the flow rate or air volume must have caused insufficient turbulence levels and particle impaction velocities which have significantly reduced the capsule emptying and consequently the FPFrecovered. Studies are ongoing to overcome this limitation by enlarging the piercing pin size to improve the powder emptying and dispersion.

243

In summary, the 1pin_HR_00 device is capable of aerosolising colistin and rifapentine up to

100 mg in a single capsule. This negates the need to load multiple capsules and can improve patient compliance as they do not have to remember how many capsules they have inhaled.

The device is more patient friendly compared to other commercial products with a simple instruction of “open-load-inhale”. It is also relatively easy to check the remaining dose by simply opening the mouthpiece of the device and visually inspecting the capsule. Furthermore, the simple design of the device makes it cost-effective and easier for quality control in mass production. However, there are also possible drawbacks of this high payload design. (1) If the device used for multiple doses, the drug retention/accumulation in the device could adversely affect the spinning of the capsule or ‘slough off’ the performance and impart high variability between doses. (2) Patients need to inhale multiple times to empty the capsule which may cause fatigue during consecutive inhalations.

244

5. CONCLUSION

We have shown that the modified Aerolizer® (1pin_HR_00) as a low-cost generic device can efficiently aerosolise powders with different physiochemical properties up to a maximum dose of 100 mg per capsule. This device can also perform well at the inhalation conditions achievable by CF patients. Therefore, using this device can prevent the need for multiple capsule loadings and ease the administration of high-dose medications, and this has the opportunity to enhance adherence to the inhaled regimen. Further optimisation of the device using different capsule aperture sizes to deliver a wider range of high-dose formulations for treatment of lung infections is currently underway.

245

ACKNOWLEDGEMENT

Thaigarajan Parumasivam is a recipient of the Malaysian Government Scholarship. Sharon

Leung is a research fellow supported by the University of Sydney. The authors wish to acknowledge the technical support of Australian Microscopy & Microanalysis Research

Facility at the Australian Centre for Microscopy and Microanalysis, The University of Sydney.

The work was supported by the Australian Research Council's Discovery Project

(DP150103953), the National Health and Medical Research Council Centre of Research

Excellence in Tuberculosis Control (APP1043225) and Project grant (APP104434), and the

NSW Government Infrastructure Grant to the Centenary Institute.

246

REFERENCES

[1] P. Agent, H. Parrott, Inhaled therapy in cystic fibrosis: agents, devices and regimens, Breathe, 11 (2015) 110-118. [2] M. González-Juarrero, M.P. O’Sullivan, Optimization of inhaled therapies for tuberculosis: The role of macrophages and dendritic cells, Tuberculosis, 91 (2011) 86-92. [3] R. Karbasi-Afshar, J. Aslani, M. Ghanei, Efficacy and safety of inhaler steroids in COPD patients: systematic review and meta-analysis of randomized placebo-controlled trials, Caspian J Intern Med, 5 (2014) 130-136. [4] N.R. Labiris, M.B. Dolovich, Pulmonary drug delivery. Part II: The role of inhalant delivery devices and drug formulations in therapeutic effectiveness of aerosolized medications, Br J Clin Pharmacol, 56 (2003) 600-612. [5] D.E. Geller, Comparing clinical features of the nebulizer, metered-dose inhaler, and dry powder inhaler, Respir Care, 50 (2005) 1313-1322. [6] M. Zeidler, J. Corren, Hydrofluoroalkane formulations of inhaled corticosteroids for the treatment of asthma, Treat Respir Med, 3 (2012) 35-44. [7] N. Islam, E. Gladki, Dry powder inhalers (DPIs) - a review of device reliability and innovation, Int J Pharm, 360 (2008) 1-11. [8] S. Agarkhedkar, P.S. Kulkarni, S. Winston, R. Sievers, R.M. Dhere, B. Gunale, K. Powell, P.A. Rota, M. Papania, Safety and immunogenicity of dry powder measles vaccine administered by inhalation: a randomized controlled phase I clinical trial, Vaccine, 32 (2014) 6791-6797. [9] A.S. Dharmadhikari, M. Kabadi, B. Gerety, A.J. Hickey, P.B. Fourie, E. Nardell, Phase I, single-dose, dose-escalating study of inhaled dry powder capreomycin: a new approach to therapy of drug-resistant tuberculosis, Antimicrob Agents Chemother, 57 (2013) 2613-2319. [10] Novartis Pharmaceuticals UK Ltd, TOBI Podhaler 28 mg inhalation powder, hard capsules, in, Data Pharm© 2015, http://www.medicines.org.uk/emc/medicine/24989/SPC/TOBI+Podhaler+28+mg+inhalation +powder,+hard+capsules, 2015. [11] M.R. Elkins, P. Robinson, S.D. Anderson, C.P. Perry, E. Daviskas, B. Charlton, Inspiratory flows and volumes in subjects with cystic fibrosis using a new dry powder inhaler device, Open Respir Med J, 8 (2014) 1-7. [12] R. Hodder, D. Price, Patient preferences for inhaler devices in chronic obstructive pulmonary disease: experience with Respimat(®) Soft Mist(™) Inhaler, Int J Chron Obstruct Pulmon Dis, 4 (2009) 381-390. [13] A.C. Moore, S. Stone, Meeting the needs of patients with COPD: patients' preference for the Diskus inhaler compared with the Handihaler, Int J Clin Pract, 58 (2004) 444-450. [14] P.M. Young, J. Crapper, G. Philips, K. Sharma, H.-K. Chan, D. Traini, Overcoming dose limitations using the Orbital® multi-breath dry powder inhaler, J Aerosol Med Pulm Drug Deliv, 27 (2013) 138-147. [15] P.M. Young, R.O. Salama, B. Zhu, G. Phillips, J. Crapper, H.-K. Chan, D. Traini, Multi- breath dry powder inhaler for delivery of cohesive powders in the treatment of bronchiectasis, Drug Dev Ind Pharm, 41 (2014) 859-865. 247

[16] Q. Zhou, Z. Tong, P. Tang, M. Citterio, R. Yang, H.-K. Chan, Effect of device design on the aerosolization of a carrier-based dry powder inhaler-a case study on Aerolizer(®) Foradile (®), AAPS J, 15 (2013) 511-522. [17] M. Coates, H.-K. Chan, D. Fletcher, J. Raper, Influence of air flow on the performance of a dry powder inhaler using computational and experimental analyses, Pharm Res, 22 (2005) 1445-1453. [18] M. Coates, D. Fletcher, H.-K. Chan, J. Raper, The role of capsule on the performance of a dry powder inhaler using computational and experimental analyses, Pharm Res, 22 (2005) 923-932. [19] N.Y.K. Chew, H.-K. Chan, D.F. Bagster, J. Mukhraiya, Characterization of pharmaceutical powder inhalers: estimation of energy input for powder dispersion and effect of capsule device configuration, J Aerosol Sci, 33 (2002) 999-1008. [20] M.S. Coates, H.-K. Chan, D.F. Fletcher, J.A. Raper, Effect of design on the performance of a dry powder inhaler using computational fluid dynamics. Part 2: air inlet size, J Pharm Sci, 95 (2006) 1382-1392. [21] W.H. Wong, J. Crapper, H.-K. Chan, D. Traini, P. Young, Pharmacopeial methodologies for determining aerodynamic mass distributions of ultra-high dose inhaler medicines, J Pharm Biomed Anal, 51 (2010) 853-857. [22] M. Gurjar, Colistin for lung infection: an update, J Intensive Care, 3 (2015) 3. [23] J.G.Y. Chan, X. Bai, D. Traini, An update on the use of rifapentine for tuberculosis therapy, Expert Opin Drug Deliv, 11 (2014) 421-431. [24] K. McKeage, Tobramycin inhalation powder: a review of its use in the treatment of chronic Pseudomonas aeruginosa infection in patients with cystic fibrosis, Drugs, 73 (2013) 1815-1827. [25] P. Selvam, D. McNair, R. Truman, H.D.C. Smyth, A novel dry powder inhaler: effect of device design on dispersion performance, Int J Pharm, 401 (2010) 1-6. [26] Q. Zhou, D. Morton, H. Yu, J. Jacob, J. Wang, J. Li, H.-K. Chan, Colistin powders with high aerosolisation efficiency for respiratory infection: preparation and in vitro evaluation, Immunol Rev, 102 (2013) 3736-3747. [27] J.G.Y. Chan, C.C. Duke, H.X. Ong, J.C.Y. Chan, A.S. Tyne, H.-K. Chan, W.J. Britton, P.M. Young, D. Traini, A novel inhalable form of rifapentine, J Pharm Sci, 103 (2014) 1411- 1421. [28] ChemNet, 32986-56-4 tobramycin, in: C.C.N.a.C.c.a.s.o.Z.N.C. Global Chemical Exchange, Ltd. (Ed.) http://www.chemnet.com/cas/en/32986-56-4/Tobramycin-Base.html, 2016. [29] H.A. Tiddens, D.E. Geller, P. Challoner, R.J. Speirs, K.C. Kesser, S.E. Overbeek, D. Humble, S.B. Shrewsbury, T.A. Standaert, Effect of dry powder inhaler resistance on the inspiratory flow rates and volumes of cystic fibrosis patients of six years and older, J Aerosol Med, 19 (2006) 456-465. [30] U.S. Pharmacopeial Convention, Tobamycin inhalation solution in: The tobramycin inhalation solution revision bulletin The United States Pharmacopeial Convention, http://www.usp.org/usp-nf/official-text/revision-bulletins/tobramycin-inhalation-solution, 2015.

248

[31] D.P. Miller, T. Tan, T.E. Tarara, J. Nakamura, R.J. Malcolmson, J.G. Weers, Physical Characterization of Tobramycin Inhalation Powder: I. Rational Design of a Stable Engineered- Particle Formulation for Delivery to the Lungs, Mol Pharm, 12 (2015) 2582-2593. [32] A.K. Dash, R. Suryanayanan, Solid-state properties of tobramycin, Pharm Res, 8 (1991) 1159-1165. [33] T. Srichana, G.P. Martin, C. Marriott, Dry powder inhalers: the influence of device resistance and powder formulation on drug and lactose deposition in vitro, Eur J Pharm Sci, 7 (1998) 73-80. [34] R.W. Dal Negro, Dry powder inhalers and the right things to remember: a concept review, Multidiscip Respir Med, 10 (2015) 13. [35] A.R. Clark, A.M. Hollingworth, The relationship between powder inhaler resistance and peak inspiratory conditions in healthy volunteers - implications for in vitro testing, J Aerosol Sci, 6 (1993) 99-110. [36] M. Kruse, B. Rosenkranz, C. Dobson, G. Ayre, A. Horowitz, Safety and tolerability of high-dose formoterol (Aerolizer) and salbutamol (pMDI) in patients with mild/moderate, persistent asthma, Pulm Pharmacol Ther, 18 (2005) 229-234. [37] W.P. Adams, S.L. Lee, R. Plourde, R.A. Lionberger, C.M. Bertha, W.H. Doub, J.-M. Bovet, A.J. Hickey, Effects of device and formulation on in vitro performance of dry powder inhalers, AAPS J, 14 (2012) 400-409. [38] M.S. Hassan, R.W.M. Lau, Effect of particle shape on dry particle inhalation: study of flowability, aerosolization, and deposition properties, AAPS PharmSciTech, 10 (2009) 1252- 1262. [39] R. Vehring, Pharmaceutical particle engineering via spray drying, Pharm Res, 25 (2008) 999-1022. [40] J. Chan, A. Tyne, A. Pang, H.-K. Chan, P. Young, W. Britton, C. Duke, D. Traini, A rifapentine-containing inhaled triple antibiotic formulation for rapid treatment of tubercular infection, Pharm Res, 31 (2013) 1239-1253. [41] R. Vanbever, J.D. Mintzes, J. Wang, J. Nice, D. Chen, R. Batycky, R. Langer, D.A. Edwards, Formulation and physical characterization of large porous particles for inhalation, Pharm Res, 16 (1999) 1735-1742. [42] Z.B. Tong, B. Zheng, R.Y. Yang, A.B. Yu, H.K. Chan, CFD-DEM investigation of the dispersion mechanisms in commercial dry powder inhalers, Powder Technol, 240 (2013) 19- 24. [43] T. Kondo, T. Tanigaki, M. Hibino, M. Ohe, S. Kato, Resistances of dry powder inhalers and training whistles and their clinical significance, Arerugi, 63 (2014) 1325-1329. [44] F. Lavorini, The challenge of delivering therapeutic aerosols to asthma patients, ISRN Allergy, 2013 (2013) 17.

249

CHAPTER 6

INHALATION OF RESPIRABLE CRYSTALLINE

RIFAPENTINE PARTICLES INDUCES PULMONARY

INFLAMMATION

This chapter has been published in Molecular Pharmaceutics, 2017;14: 328-335, with the title

‘Inhalation of respirable crystalline rifapentine particles induces pulmonary inflammation’.

Authors: Parumasivam T, Ashhurst AS, Nagalingam G, Britton WJ, Chan H-K.

250

ABSTRACT

Rifapentine is an anti-tuberculosis (TB) drug with a prolonged half-life, but oral delivery results in low concentrations in the lungs because of its high binding (98 %) to plasma proteins.

We have shown that inhalation of crystalline rifapentine overcomes the limitations of oral delivery by significantly enhancing and prolonging the drug concentration in the lungs. The delivery of crystalline particles to the lungs may promote inflammation. This in vivo study characterises the inflammatory response caused by pulmonary deposition of the rifapentine particles. The rifapentine powder was delivered to BALB/c mice by intratracheal insufflation at a dose of 20 mg/kg. The inflammatory response in the lungs and bronchoalveolar lavage

(BAL) was examined at 12 h, 24 h and 7 days post-treatment by flow cytometry and histopathology. At 12 h and 24 h post-treatment, there was a significant influx of neutrophils into the lungs, and this returned to normal by day 7. A significant recruitment of macrophages occurred in the BAL at 24 h. Consistent with these findings, histopathological analysis demonstrated pulmonary vascular congestion and significant macrophage recruitment at 12 h and 24 h post-treatment. In conclusion, the pulmonary delivery of crystalline rifapentine caused a transient neutrophil-associated inflammatory response in the lungs that resolved over 7 days.

This observation may limit pulmonary delivery of rifapentine to once a week at a dose of 20 mg/kg or less. The effectiveness of weekly dosing with inhalable rifapentine will be assessed in murine Mycobacterium tuberculosis infection.

251

1. INTRODUCTION

Tuberculosis (TB) is a leading cause of death globally and is predominantly a respiratory infectious disease caused by Mycobacterium tuberculosis. It is estimated that one-third of the world’s population has been exposed to M. tuberculosis with almost 9.6 million new cases and

1.5 million deaths annually [1]. The current treatment of TB is based on oral administration of at least three first-line drugs in combination (i.e. isoniazid, rifampicin, ethambutol, pyrazinamide) with direct observation of patient compliance. This strategy has been widely promoted and implemented by the World Health Organization (WHO) as DOTS (directly observed treatment, short course) [2]. However, the treatment is challenging owing to the requirement of at least 6 months therapy with high dose antibiotics and the associated side effects that often reduce patient compliance and treatment success.

Inhalation drug delivery, following the natural route of infection, is a rational approach to improve the current anti-TB regimen. The advantages of inhalation drug delivery are: (1) it delivers the drug directly to the lungs that results in a higher local drug concentration than can be achieved by oral or parenteral administration [3, 4], (2) it reduces the systemic exposure of the drug [5], and (3) it avoids degradation of the drug by the liver and the acidic gastric environment [6]. These features of pulmonary delivery can reduce the drug dose and adverse effects, and therefore has the potential to shorten the treatment duration.

Rifapentine is an US Food and Drug Administration (FDA) approved drug administered intermittently for treatment of latent TB. Following oral administration, however, rifapentine fails to reach effective concentrations in the lungs owing to high plasma protein binding (98

%)[7]. In addition, the bioavailability of the drug is highly influenced by concomitant food

252

intake and induction of its own metabolism [8, 9]. Studies have shown that these barriers can be overcome if rifapentine is delivered directly to the lungs [10, 11]. In a previous study, we demonstrated that the intratracheal delivery of inhalable crystalline rifapentine to mice at a dose of 20 mg/kg produced a lung Cmax 100-fold higher than the intraperitoneal administration at same dose [10]. Furthermore, the lung Tmax was extended to 2 h when the drug was given by the intratracheal route compared to 1 h after intraperitoneal delivery [10]. The blood plasma and lung drug concentrations were also maintained above the minimum inhibitory concentration (MIC) against M. tuberculosis (0.03 – 0.06 µg/mL) [12] for more than 24 h post- treatment [10]. These findings suggest that the inhalable crystalline rifapentine significantly enhances and prolongs the drug concentration in the lungs, and this has the potential to reduce the therapeutic dose and the treatment duration whilst maintaining systemic concentrations.

Inhalation therapy is now accepted treatment for a number of respiratory diseases, such as asthma and cystic fibrosis,[13] but no inhaled medication has been approved for the treatment of TB. One limitation on the use of inhalable antibiotics for treatment of TB is the potential to exacerbate inflammation in the lungs. Therefore, in this report, we have examined the effects on pulmonary inflammation and cellular influx after intratracheal insufflation of respirable- size crystalline rifapentine particles at a dose of 20 mg/kg in mice. We report the changes in the leukocyte populations in the lungs and the bronchoalveolar lavage (BAL), and the histopathological changes in the lungs after drug exposure. Although the cytotoxicity of rifapentine on in vitro respiratory cells has been documented [14-17], this is the first study to characterise in vivo the effects of pulmonary delivery of respirable-size rifapentine crystals.

This is a necessary foundation prior to proceeding using the rifapentine in future murine TB efficacy studies and human clinical trials. Furthermore, this study contributes to better understand the safety of inhaled therapies for TB.

253

2. MATERIALS AND METHODS

Rifapentine dry powder production and characterisation

The rifapentine powder was produced following a method detailed in Chan et al. [17]. Briefly, rifapentine (Hangzhou ICH Imp & Exp Co. Ltd., Hangzhou, China) was dissolved in a co- solvent system consisting of acetone (Ajax Finechem Pty Ltd., NSW, Australia) and

UltraPureTM DNase/RNase-free distilled water (Thermo Fisher Scientific, NSW, Australia)

(1:1, v/v) at a concentration of 4 mg/mL. The system was then heated at 67°C to evaporate the acetone and formed a crystalline suspension. The suspension was homogenised at 10 000 rpm for 2 minutes (Silverson L4RT High Shear Mixer, Silverson, Chesham, UK) and spray dried using a B-290 mini spray dryer (Buchi Labortechnik AG, Flawil, Switzerland) in an open-loop configuration using the following parameters: inlet temperature of 60 °C, atomiser of 60 mm, aspirator of 100 %, and feed rate of 2 mL/min.

The morphology of the particles was examined using scanning electron microscopy (SEM)

(FESEM, Hitachi S4500, Japan) at an acceleration voltage of 5 kV. The volume equivalent diameters of the particles were measured using laser diffraction coupled with a Scirocco dry powder disperser (Mastersizer 2000, Malvern Instruments, Worcestershire, UK) at a refractive index (RI) of 1.63 and dispersion air pressure of 3.5 bars.

The aerodynamic properties of the powder was tested at flow rates of 60 L/min and 100 L/min for 4.0 secs and 2.4 secs, respectively with both representing an air volume of 4.0 L. A capsule

(size 3, Capsugel, NSW, Australia) containing 10 ± 0.5 mg powder was loaded into an

Osmohaler® and actuated through a multi-stage liquid impinger (MSLI) (Copley, Nottingham,

254

UK) to simulate an inspiration. The dispersion was performed in triplicate in a climate controlled cabinet: temperature, 20 ± 5°C and relative humidity (RH), 50 ± 3 %.

Methanol:water (3:1, v/v) with 2 % ascorbic acid was used as a rinsing solvent. The drug mass was quantified using high-performance liquid chromatography (HPLC) [16, 17]. The mobile phase consisted of 50 mM phosphate buffer (pH 3.0):acetonitrile:tetrahydrofuran at a ratio of

53:42:5 (%, v/v), and the stationary phase was a µBondapakTM C18 (3.9 × 300 mm2) column

(Waters, Milford, Massachusetts). The system (Model LC-20, Shimadzu, Kyoto, Japan) was operated at an isocratic flow rate of 1 mL/min for 10 min. UV detection was performed at a wavelength of 250 nm (SPD-20A UV/VIS detector, Shimadzu) and the injection volume was

20 µL. A calibration curve with a R2 value of 1.0 was obtained with standard concentrations ranging from 0.001 to 0.1 mg/mL. The emitted dose was calculated as the amount of drug exited the device and capsule relative to the total drug recovery from HPLC. The recovered fine particle fraction (FPFrecovered) was represented as the percentage of particles with an aerodynamic diameter less than 5 µm relative to the total drug recovery from HPLC. Mass median aerodynamic diameter (MMAD) and geometric standard deviation (GSD) were calculated from the log-probability plot of the MSLI data, following Appendix XII C of the

British Pharmacopoeia.

Mice and ethics statement

Female 6 - 10 week old BALB/c mice were obtained from Animal BioResources (Moss Vale,

NSW, Australia) and were housed in specific-pathogen-free facilities at the Centenary Institute,

Camperdown, NSW, Australia. The experiment was conducted in accordance with the

Australian code of practice for the care and use of animals for scientific purpose and approved

255

by the Sydney Local Health District Animal Welfare Committee (protocol number

2014/031A).

Intratracheal delivery of rifapentine in a murine model

Rifapentine was administered to mice via insufflation (DP-4M, Penn-Century, Philadelphia) at a well-tolerated oral dose used in mice and humans (20 mg/kg) [9, 18]. The mouse was anaesthesised with ketamine/xylazine (90/10 mg/kg) by intraperitoneal injection and secured on an intubation platform. The insufflator was then inserted into the trachea until the tip was immediately adjacent to the first bifurcation and the drug directly aerosolised onto the lung surface. The insufflator was weighed before and after the aerosolisation to measure the delivered dose. The mouse was then injected intraperitoneally with atipamezole (0.0125 mg) and rested on a heating pad and monitored until self-righting. Mice were euthanised at 12 h, 24 h and 7 days post-treatment by CO2 asphyxiation, and lungs (perfused with PBS + 20 U/mL

Heparin) and BAL were collected. Three lobes of the lungs and BAL were processed for total cell counts and immune cell characterisation via flow cytometry. One lobe of the lungs was used for histopathological analysis. Another one lobe was used for drug quantification (data not reported). The experiments were performed independently twice.

The control mice were untreated (Nil) and did not undergo insufflations to determine basal values. Morello et al. reported that insufflation (air only) in female BALB/c mice resulted only in a minor congestion and hemorrhage, and resolved over time [19]. Similarly, in the present study, no macroscopic indication of significant trauma was observed in the lungs of the insufflated/treated mice. In addition, there was no significant increase in total leukocyte counts

256

in the lungs or BAL of the treated mice compared to the untreated group. Therefore, the inflammation caused by insufflation alone is likely to be negligible.

Flow cytometry analysis

Lung tissue was digested in RPMI 1640 media (Bacto Laboratories, Mt. Pritchard, NSW,

Australia) containing fetal calf serum (FCS) (10 %, v/v), collagenase IV (50 U/mL, Sigma-

Aldrich, Castle Hill, Australia) and DNAse I (13 µG/mL, Sigma-Aldrich, Castle Hill,

Australia) for 30 min at 37°C. After processing the lung tissue and BAL to single cell suspensions, erythrocytes were removed utilising ammonium-chloride-potassium (ACK) lysis buffer and the cells were then re-suspended in complete media. Fc receptors were blocked by incubation of cell suspensions with anti-CD16/CD32 (BD Pharmingen, San Jose, CA) in FACS buffer (PBS with 2 % FCS), followed by antibody staining for major leukocyte subsets.

Antibody panel utilised: anti-CD11b AF700 (M1/70, BD Pharmingen, San Jose, CA), anti-

CD11c PECy7 (HL3, BD Pharmingen, San Jose, CA), anti-F4/80 PE (T45-2342, BD

Pharmingen, San Jose, CA), anti-MHCII APC (M5/114.15.2, eBioscience, San Diego, CA), anti-Ly6G PerCPCy5.5 (1A8, Biolegend, San Diego, CA), anti-B220 APCCy7 (RA3-6B2, BD

Pharmingen, San Jose, CA), anti-CD45.2 PB (104, Biolegend, San Diego, CA), anti-CD3 FITC

(145-2C11, BD Pharmingen, San Jose, CA) and live/dead fixable blue dead cell stain

(Invitrogen, Carlsbad, CA). Data were acquired on the BD LSRFortessa and analysed with

FlowJo (TreeStar, Ashland, OR).

257

Histopathology examination

The lung lobes designated for histopathology were inflated and immersed in 10 % neutral buffered formalin, and stained with hematoxylin and eosin (H & E). Histopathological evaluation was measured by microscopy on 5 µm sections of the stained lung. Inflammation was scored by two independent histopathologists who were blinded to the treatment conditions.

Statistical analysis

Data were expressed as mean ± standard deviation (SD) or standard error of the mean (SEM) as specified. Differences between treatment groups were evaluated by ANOVA with the

Bonferroni post hoc comparison (GraphPad Prism 6 software), and p < 0.05 was considered statistically significant.

258

3. RESULTS

Rifapentine powder characterisation

The rifapentine particles were elongated crystals with physical width × length of at least 0.5 ×

1.0 µm as shown in the scanning electron micrograph (Figure 1), similar to those reported in our previous study [15]. The volume median diameter (D50) of the particles measured by laser diffraction was 1.47 ± 0.12 µm with a span of 1.56 ± 0.09. When the powder was dispersed at flow rates of 60 L/min and 100 L/min with an air volume of 4.0 L, the fine particle fraction

(FPF < 5 µm) was 69.5 ± 1.3 % and 78.4 ± 1.1 %, respectively (Table 1). This was corresponding to a MMAD of 2.21 ± 0.1 µm and 1.85 ± 0.1 µm, respectively. The GSD was indifferent at both conditions with values between 1.75 and 1.78. Irrespective of the flow rates, the drug mostly deposited in stages 3 and 4, and filter stage (Figure 2). However, significantly higher device retention was observed at 60 L/min (22.3 ± 1.3 %) than 100 L/min (12.4 ± 3.3

%).

Figure 1. Elongated crystalline rifapentine particles viewed under scanning electron

microscope.

259

40 60 L/min

100 L/min 30

Recovered mass(%) Recovered 10

0 Capsule Device Adaptor Throat Stage 1 Stage 2 Stage 3 Stage 4 Filter (10.07 µm) (5.27 µm) (2.40 µm) (1.32 µm) (<1.32µm)

Figure 2. In vitro deposition profile of the rifapentine following MSLI dispersion at a flow rate of 60 L/min and 100 L/min with an air volume of 4.0 L (n = 3).

Table 1. Emitted dose, fine particle fraction (FPF < 5µm), mass median aerodynamic diameter

(MMAD), and geometric standard deviation (GSD) for rifapentine powder dispersed at 60

L/min and 100 L/mina.

Flow rate 60 L/min 100 L/min

Emitted dose (%) 73.8 ± 0.8 86.6 ± 2.5

FPF < 5 µm (%) 69.5 ± 1.2 78.4 ± 0.2

MMAD (µm) 2.21 ± 0.1 1.85 ± 0.1

GSD 1.78 ± 0.1 1.75 ± 0.1 amean ± SD, n = 3

260

Flow cytometry analysis

A significant increase in neutrophils was observed in the lungs at 12 h and 24 h after treatment

(Figure 3), an increase of approximately 20 % and 10 %, respectively compared to the untreated control (p < 0.01). However, the neutrophil population returned to normal by day 7. There were no significant differences in the other leukocyte populations in the lungs, except there was a trend for increase in macrophage/monocyte populations at 12 h and 24 h (p > 0.05), but this also returned to baseline by day 7.

Similar cellular changes were also noted in the BAL after treatment (Figure 4). At 12 h and 24 h post-dosing, there was a significant recruitment of neutrophils so that they comprised 80 % and 60 %, respectively of all leukocytes in the BAL (p < 0.05). Again, neutrophil counts returned to baseline by 7 days. There was a significant increase in the macrophage/monocyte and alveolar macrophage populations in the BAL at 24 h (p < 0.05), along with a small reduction in the T-cell, B-cell and dendritic cell (DC) populations at 12 h and 24 h post-dosing

(p > 0.05).

In summary, pulmonary delivery of rifapentine caused a transient neutrophil-associated inflammation in the lungs that had resolved over 7 days. Please refer to the Supporting

Information for a detailed analysis of immune cell populations in the lungs and BAL (Appendix

5).

261

Nil 12h 24h 7d

NeutroNephilsutrophils B-cellsB-cells DC DC Mac/mono AlveolMaar mcacr/mopohnagoes pDC Alveolar macrophages T-cells Other pDC T-cells Other

B C D

6 5 1.5×10 2.5×104 6×10

** * s e

4 g

2.0×10 a

h s

l 6

p 5 i

1.0×10 o 4×10

h n

4 r p

o 1.5×10

m u

a r e 5 1.0×10 5

5.0×10 a M

N 2×10

l o

5.0×103 e

l A 0.0 0.0 0 Nil 12h 24h 7d Nil 12h 24h 7d Nil 12h 24h 7d Figure 3. Enumeration of leukocyte (CD45.2+) populations in the lungs of untreated mice, or at 12 h, 24 h and 7 days post-treatment. (A) Frequency of populations as a proportion of total lung leukocytes. Neutrophils (Ly6G+ CD11b+), B-cells (B220+ CD11c-), dendritic cells (DC) (B220- F4/80- CD11chi), combined macrophage/monocyte populations (Mac/mono) (B220- F4/80- CD11b+ CD11c-), alveolar macrophages (B220- F4/80+ CD11b+ CD11c-), plasmacytoid dendritic cells (pDC) (B220+ CD11c+), T-cells (CD3+) and others (unidentified cells). Enumeration of (B) neutrophils, (C) combined 262 macrophage/monocyte populations (Mac/mono), and (D) alveolar macrophages. Differences between treatment groups were evaluated by ANOVA with

Bonferroni post hoc comparison (*p < 0.05, **p < 0.01).

Nil 12h 24h 7d

Neutrophils B-cells DC Mac/mono Alveolar macrophages pDC T-cells Other

B C D

* 5×104 ** * 6 * * 1×10 2.5×105 *

4 s

4×10 e 5

8×10 g 2.0×105

n l

4 p i

o 3×10 o h 5 5

6×10 r

p 1.5×10

r c

t 4

a m u 5 2×10 5

4×10 1.0×10

N l

5 4 o

2×10 1×10 e 5.0×104

l A 0 0 0.0 Nil 12h 24h 7d Nil 12h 24h 7d Nil 12h 24h 7d

Figure 4. Enumeration of leukocyte (CD45.2+) populations in the BAL of untreated mice, or at 12 h, 24 h and 7 days post-treatment. (A) Frequency of populations as a proportion of total leukocytes in the BAL. Neutrophils (Ly6G+ CD11b+), B-cells (B220+ CD11c-), dendritic cells (DC) (B220- F4/80-

CD11chi), combined macrophage/monocyte populations (Mac/mono) (B220- F4/80- CD11b+ CD11c-), alveolar macrophages (B220- F4/80+ CD11b+

CD11c-), plasmacytoid dendritic cells (pDC) (B220+ CD11c+), T-cells (CD3+) and others (unidentified cells). Enumeration of (B) neutrophils, (C) 263 combined macrophage/monocyte populations (Mac/mono), and (D) alveolar macrophages. Differences between treatment groups were evaluated by

ANOVA with Bonferroni post hoc comparison (*p < 0.05, **p < 0.01).

Histopathology examination

Selected photomicrographs from the lungs of mice are depicted in Figure 5 and the interpretations are summarised in Table 2. When compared to control mice (Figure 5A), lungs of mice taken at 12 h after treatment showed a moderate cytoplasmic vacuolation and blebbing of the respiratory epithelium that suggested an epithelial cell degeneration (Figure 5B).

Furthermore, mild necrosis occurred in the large to small bronchioles on the epithelium lining with mucin and karyorrhectic particles adhered to these small eroded surfaces. A mild perivascular inflammation characterised by neutrophil influx was also observed (Figure 5C).

The examination at 24 h post-treatment (Figure 5D) showed a severe congestion in the parenchyma, compared to that of observed in the control mice or 12 h post-treatment. Extra- and intra-pulmonary airways including the terminal bronchiole displayed cytoplasmic vacuolation in at least 50 % of the epithelial cells and nearly one-third of the vacuole was mainly solitary. Moreover, the epithelial cells had shed and/or proliferated in small bronchioles with mononuclear cells in the peribronchial zone. In some airways, particularly in small bronchioles, a focal increase of epithelial cells with smaller nuclear to cytoplasmic (NC ratio) was present, suggesting early regeneration of cells. A significant recruitment of alveolar macrophages and edema was also observed (Figure 5E).

The day 7 post-treatment (Figure 5F) observations showed that at least 60 % of the parenchyma was conspicuously congested compared to the untreated lung. The vascular outline of the alveolar walls was enhanced by proteinaceous deposits that obscured capillary and its lumen.

The clumps of granular proteinaceous material also occupied some alveolar sacs together with few attached mononuclear and macrophage cells.

264

In summary, a prominent vacuolar cytoplasmic degenerative change occurred at 12 h that had intensified by 24 h. In addition, at 24 h a minor focal loss of bronchial epithelium was accompanied by an early stage of regenerative hyperplasia. This vascular congestion continued up to day 7. Nonetheless, no significant pleural changes were observed.

265

A B

C D

E F

Figure 5. Photomicrographs of hematoxylin and eosin-stained (H & E) lung sections. (A) Lung section of an untreated mouse (Nil) (magnification x 10). (B) Lung section examined at 12 h after treatment showed a micro-vacuolar cytoplasmic change characterised by congestion

(black arrow) and blebbing (grey arrow) of the epithelial lining (magnification x 10). (C) At 12

266

h post-treatment, perivascular inflammation characterised by neutrophil influx (blue arrow)

was also observed (magnification x 10). (D) At 24 h post treatment, lung section showed

intense congestion (black arrow) when compared to the control and 12 h-post dosing group

(magnification x 10), and (E) recruitment of alveolar macrophages and occurrence of edema

(grey arrow) (magnification x 20). (F) Day 7 lung section showed a presence of alveolar

macrophages, granulomatous inflammation and alveolar vascular congestion (magnification x

20).

Table 2. Summary of histopathological findings in the lungs of animals received crystalline

rifapentine.

Parameters Control 12 h 24 h Day 7

Alveolar congestion - + ++ ++

Vacuolisation - + + +

Presence of neutrophils - + ++ ++

Presence of monocytes or macrophages - + ++ ++

Necrosis of epithelial - + + -

Edema - - + +

Pleural change - - - -

Abbreviation: (-), within normal limits; (+), mild; (++), severe.

267

4. DISCUSSION

The rifapentine powder was found suitable for aerosolisation with a high FPF between 69.5 % and 78.4 %. The elongated morphology of the particles reduced the cohesive van der Waals forces and deposited via interception for a high respirable fraction [17]. This is also reflected in a minimal deposition in the throat and stage 1, and higher retention in stage 3 to filter stage of the MSLI. However, the FPF produced at 100 L/min (78.4 ± 0.2 %) was significantly higher compared to when dispersed at 60 L/min (69.5 ± 1.2 %). The inferior FPF was due to insufficient powder emptying from the device at 60 L/min, suggesting the elongated particles required high energy input to disperse the powder as revealed by the marked pressure drop at

100 L/min (~4.0 kPa) than that of 60 L/min (~1.4 kPa). Nonetheless, this high device retention in clinical application is likely to be optimised by multiple breathing inspirations. Thus, the crystalline rifapentine demonstrated in vitro aerosol properties suitable for inhaled administration.

In our previous studies, we have established that solubilised rifapentine was tolerated by human airway epithelial (Calu-3), human alveolar basal epithelial (A549) and human THP-1 monocyte-derived macrophage cell lines up to a concentration of 60 µg/mL [14, 16, 17, 20]. In contrast, tobramycin, a currently FDA-approved drug for treatment of cystic fibrosis via the inhalation route, is non-toxic to lung epithelial cells up to a concentration of 1.0 g/mL [21].

However, the cytotoxicity of rifapentine could be overestimated in the in vitro assays as the drug was dissolved in an organic solvent (i.e. DMSO) prior to use in the assays, while a gradual dissolution is expected after deposited in the lungs. The in vitro assays also differ from the actual characteristics of in vivo delivery to the lungs, such as mucociliary clearance of the pulmonary region that will have a significant impact on the particle dissolution and retention

268

in the lungs. Therefore, assessing the impact of pulmonary delivery in vivo is essential for determining a rational dosage regimen for further preclinical and clinical studies.

A single dose pulmonary insufflation of rifapentine caused a neutrophilic inflammation in the lungs at 12 h and 24 h post-exposure that resolved by 7 days. Neutrophils are the first leukocytes recruited to phagocytose and denature foreign materials deposited in the lungs through the release of reactive oxygen species, defensins, lactoferrin, cathepsins and lysozymes

[22, 23]. Increased secretion of reactive oxygen species, such as superoxide anion and proteolytic enzymes (e.g. neutrophil elastase) can cause significant pulmonary inflammation and lung damage. The physiological effect of the inflammation caused by the neutrophil influx in the present study is unknown. None of the mice were distressed, suggesting the insufflation of rifapentine had not caused any acute lung damage for 7 days after the treatment. The histology further revealed that the lung damage was reversible with evidence of some regeneration of epithelial cells and no significant pleural changes after 24 h and 7 days post- treatment. Similar results were reported when a single dose of a cationic lipid suspension was administered to BALB/c mice by nasal instillation [24]. This caused a significant influx of neutrophils into the lungs and BAL at 1 to 3 days that returned to baseline by 14 days. There were also increased levels of the pro-inflammatory cytokines, interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α) and interferon-γ (IFN- γ) that peaked at 1 to 2 day post-treatment and returned to baseline by 14 days [24]. On the other hand, Patil-Gadhe et al. showed that repeated intratracheal insufflations (28 days daily) of rifapentine proliposomes at doses of 1 and 5 mg/kg in Wistar rats did not elicit pulmonary irritation, as revealed by minimal changes in neutrophil populations. However, when the multiple doses were increased to 10 mg/kg which is only 50 % of the dose used in this study, there was a significant mortality due to excessive

269

inflammatory responses in the lungs [15]. Therefore, careful attention is required if multiple doses of rifapentine at 20 mg/kg are to be delivered via the pulmonary route.

Another important finding was the increased number of alveolar macrophages (~15-20 %) and to lesser extent macrophage/monocyte populations in the BAL at 24 h post-exposure, both of which resolved by day 7. This was consistent with the lung histology analysis where the alveolar spaces were dominated by foamy macrophages at 24 h (Figure 5E). This may be due to the activation of the alveolar macrophages to phagocytose the rifapentine particles, as well as recruitment of inflammatory monocytes from the peripheral blood that differentiate into tissue macrophages [25]. Similarly, Freimark et al. showed an influx of alveolar macrophages in the lungs of C57BL/6 mice after intratracheal instillation of liposome suspensions containing an immunogenic plasmid. This resulted in enhanced immune responses, possibly through the increased cellular uptake of the particles [26]. Therefore, both alveolar and monocyte-derived macrophages may have phagocytosed the slowly dissolving rifapentine particles. This has the potential to directly target M. tuberculosis residing within macrophages and possibly shorten the duration of treatment for TB.

The DC, B-cell and T-cell counts in the BAL were slightly reduced at 12 h and 24 h time points and returned to normal by day 7, suggesting the rifapentine deposition likely affected the antigen presenting cells (APC) distribution in the lungs. A previous study demonstrated that the DCs were efficient at phagocytosing polystyrene nanoparticles (50 nm and 500 nm diameter in size) and transporting them from lung to the draining lymph node [27]. Furthermore, the inflammation in the lungs prior to the particle insufflation was found to enhance the maturation and migration of DCs, and increased the drainage of the particles via the lymphatic system [27,

28]. A similar study also reported that the footpad injection of polystyrene nanoparticles with

270

a diameter of 20 nm in diameter were taken up by DCs, macrophages and B-cells, and trafficked the particles to the lymph node [29]. These findings support the possibility that the rifapentine particles trafficked from lungs to the draining lymph nodes. The pharmacokinetic data also supports this possibility as intratracheal insufflation of rifapentine at the same dose resulted in systemic delivery of the drug with a Tmax of 12 h in the spleen [10]. Presumably, the clearance of rifapentine particles from the lungs was driven by a combination of cell-mediated transport and drainage via the lymphatic system, and also by macrophages through mucociliary clearance.

One of the critical issues in inhalation toxicology is the dose for delivery. Unlike oral and intravenous routes, it is difficult to determine the delivered dose following inhalation exposure as only a fraction of the inhaled dose will deposit in the lungs. In the present study, an intratracheal device (DP-4M® insufflator, Penn-Century) was used which has been proven to have high accuracy and reproducibility for delivery of drugs directly to the lungs [30, 31]. This delivery technique also widely distributes the powder throughout the lungs of mice, and prevents pharyngeal and upper airway losses [31]. However, one limitation of this method is that the insufflator is only efficient for delivering a minimum dose of 0.4 mg corresponding to

20 mg/kg of rifapentine. One of the approaches to overcome this barrier is to reduce the proportion of the active ingredient with a carrier, such as lactose or mannitol, in order to increase the dry powder mass loaded into the insufflator. The addition of such a carrier, however, may alter the pharmacokinetics of the drug after deposition in the lungs [32].

Furthermore, the risk of inflammation due to pulmonary delivery of rifapentine may be overestimated in this model as the drug was delivered using a DP-4M® insufflators under positive pressure which is different to aerosol deposition during tidal breathing. In the present study, 200 µL air pulses were utilised as recommended by the manufacturer [33] for delivery

271

of the powder under positive pressure. It is therefore possible that some particles may have been deposited in regions of the lungs that are not readily accessible by inhalation under ambient pressure. Hoppentocht et al. reported that an air volume of 200 µL was insufficient for aerosolisation, with 500 µL and 1000 µL air pulses providing superior deagglomeration of powders [34]. However, the maximum air volume that is safe to use for insufflation in mice is

250 µL [19]. Furthermore, during normal breathing, larger crystals would deposit in the upper airways and be eliminated before reaching the lungs. The forcible deposition of these larger crystals or agglomerates directly into the lungs by insufflations may have exacerbated the neutrophil influx. It has also been reported that even spray instillation of saline, a solution considered as “safe” and commonly used in nebulisation caused a transient neutrophil influx in rats [35]. Therefore, it would be valuable to assess the pulmonary inflammation when rifapentine is delivered via inhalation during tidal breathing in the future.

272

5. CONCLUSION

We have shown that the intratracheal delivery of respirable crystalline rifapentine at a dose of

20 mg/kg caused a neutrophil-associated transient pulmonary inflammation that resolved over

7 days period. This may limit the pulmonary delivery of rifapentine to once a week at the given dose or less. The efficacy of weekly dosing with inhalable rifapentine in a murine model of M. tuberculosis infection is ongoing. Furthermore, this finding suggests that the inhaled antibiotics may be more suitable to be administered as an adjunct with an oral regimen where patients may require intermittent dosing such as weekly, but this will depend on the powder formulation and pharmacokinetic profile of the drug.

273

ACKNOWLEDGEMENT

Thaigarajan Parumasivam and Anneliese Ashhurst are recipients of the Malaysian Government

Scholarship and Australian Postgraduate Award, respectively. The authors are thankful to Tina

Cardamone and Rolfe Howlett from Histopathology and Organ Pathology, The University of

Melbourne, Australia for the histopathological analysis. The authors acknowledge Australian

Microscopy & Microanalysis Research Facility at the Australian Centre for Microscopy and

Microanalysis, The University of Sydney for their scientific and technical assistance. The work was supported by the Australian Research Council’s Discovery Project (DP120102778), the

National Health and Medical Research Council Centre of Research Excellence in Tuberculosis

Control (APP1043225) and project grant (APP104434), and the NSW Government

Infrastructure Grant to the Centenary Institute.

274

REFERENCES

[1] World Health Organization, Global tuberculosis report 2015, in: http://apps.who.int/iris/bitstream/10665/191102/1/9789241565059_eng.pdf?ua=1 (Ed.), 2016. [2] World Health Organization, Treatment of tuberculosis: guidelines. 4th edition, 2010. [3] A. Misra, A.J. Hickey, C. Rossi, G. Borchard, H. Terada, K. Makino, P.B. Fourie, P. Colombo, Inhaled drug therapy for treatment of tuberculosis, Tuberculosis, 91 (2011) 71-81. [4] P. Muttil, J. Kaur, K. Kumar, A.B. Yadav, R. Sharma, A. Misra, Inhalable microparticles containing large payload of anti-tuberculosis drugs, Eur J Pharm Sci, 32 (2007) 140-150. [5] T. Parumasivam, R.Y.K. Chang, S. Abdelghany, T.T. Ye, W.J. Britton, H.-K. Chan, Dry powder inhalable formulations for anti-tubercular therapy, Adv Drug Deliv Rev, 102 (2016) 83-101. [6] N.R. Labiris, M.B. Dolovich, Pulmonary drug delivery. Part I: physiological factors affecting therapeutic effectiveness of aerosolized medications, Br J Clin Pharmacol, 56 (2003) 588-599. [7] K. Reith, A. Keung, P.C. Toren, L. Cheng, M.G. Eller, S.J. Weir, Disposition and metabolism of 14C-rifapentine in healthy volunteers, Drug Metab Dispos, 26 (1998) 732-738. [8] J.G.Y. Chan, X. Bai, D. Traini, An update on the use of rifapentine for tuberculosis therapy, Expert Opin Drug Deliv, 11 (2014) 421-431. [9] K.E. Dooley, E.E. Bliven-Sizemore, M. Weiner, Y. Lu, E.L. Nuermberger, W.C. Hubbard, E.J. Fuchs, M.T. Melia, W.J. Burman, S.E. Dorman, Safety and pharmacokinetics of escalating daily doses of the antituberculosis drug rifapentine in healthy volunteers, Clin Pharmacol Ther, 17 (2012) 521-525. [10] J.G.Y. Chan, A.S. Tyne, A. Pang, A.J. McLachlan, V. Perera, J.C.Y. Chan, W.J. Britton, H.K. Chan, C.C. Duke, P.M. Young, D. Traini, Murine pharmacokinetics of rifapentine delivered as an inhalable dry powder, Int J Antimicrob Agents, 43 (2014) 319-323. [11] L. Garcia Contreras, J. Sung, M. Ibrahim, K. Elbert, D. Edwards, A. Hickey, Pharmacokinetics of inhaled rifampicin porous particles for tuberculosis treatment: insight into rifampicin absorption from the lungs of guinea pigs, Mol Pharm, 12 (2015) 2642-2650. [12] N. Rastogi, K.S. Goh, M. Berchel, A. Bryskier, Activity of rifapentine and its metabolite 25-O-desacetylrifapentine compared with rifampicin and rifabutin against Mycobacterium tuberculosis, Mycobacterium africanum, Mycobacterium bovis and M. bovis BCG, J Antimicrob Chemother, 46 (2000) 565-570. [13] B.D. Kent, S.J. Lane, E.J. van Beek, J.D. Dodd, R.W. Costello, H.A.W.M. Tiddens, Asthma and cystic fibrosis: A tangled web, Pediatr Pulmonol, 49 (2014) 205-213. [14] T. Parumasivam, J.G.Y. Chan, A. Pang, D.H. Quan, J.A. Triccas, W.J. Britton, H.K. Chan, In vitro evaluation of inhalable verapamil-rifapentine particles for tuberculosis therapy, Mol Pharm, 13 (2016) 979-989. [15] A.A. Patil-Gadhe, A.Y. Kyadarkunte, M. Pereira, G. Jejurikar, M.S. Patole, A. Risbud, V.B. Pokharkar, Rifapentine-proliposomes for inhalation: in vitro and in vivo toxicity, Toxicol Int, 21 (2014) 275-282.

275

[16] T. Parumasivam, S.S.Y. Leung, D.H. Quan, J.A. Triccas, W.J. Britton, H.-K. Chan, Rifapentine-loaded PLGA microparticles for tuberculosis inhaled therapy: preparation and in vitro aerosol characterization, Eur J Pharm Sci, 88 (2016) 1-11. [17] J.G.Y. Chan, C.C. Duke, H.X. Ong, J.C.Y. Chan, A.S. Tyne, H.-K. Chan, W.J. Britton, P.M. Young, D. Traini, A novel inhalable form of rifapentine, J Pharm Sci, 103 (2014) 1411- 1421. [18] I.M. Rosenthal, M. Zhang, K.N. Williams, C.A. Peloquin, S. Tyagi, A.A. Vernon, W.R. Bishai, R.E. Chaisson, J.H. Grosset, E.L. Nuermberger, Daily dosing of rifapentine cures tuberculosis in three months or less in the murine model, PLoS Med, 4 (2007) e344. [19] M. Morello, C.L. Krone, S. Dickerson, E. Howerth, W.A. Germishuizen, Y.-L. Wong, D. Edwards, B.R. Bloom, M.K. Hondalus, Dry-powder pulmonary insufflation in the mouse for application to vaccine or drug studies, Tuberculosis, 89 (2009) 371-377. [20] T. Parumasivam, J.G.Y. Chan, A. Pang, D.H. Quan, J.A. Triccas, W.J. Britton, H.K. Chan, In vitro evaluation of novel inhalable dry powders consisting of thioridazine and rifapentine for rapid tuberculosis treatment, Eur J Pharm Biopharm, 107 (2016) 205-214. [21] G.G. Anderson, S. Moreau-Marquis, B.A. Stanton, G.A. O'Toole, In vitro analysis of tobramycin-treated Pseudomonas aeruginosa biofilms on cystic fibrosis-derived airway epithelial cells, Infect Immun, 76 (2008) 1423-1433. [22] E. Kolaczkowska, P. Kubes, Neutrophil recruitment and function in health and inflammation, Nat Rev Immunol, 13 (2013) 159-175. [23] S. Dokka, D. Toledo, X. Shi, V. Castranova, Y. Rojanasakul, Oxygen radical-mediated pulmonary toxicity induced by some cationic liposomes, Pharm Res, 17 (2000) 521-525. [24] R.K. Scheule, J.A.S. George, R.G. Bagley, J. Marshall, J.M. Kaplan, G.Y. Akita, K.X. Wang, E.R. Lee, D.J. Harris, C. Jiang, N.S. Yew, A.E. Smith, S.H. Cheng, Basis of pulmonary toxicity associated with cationic lipid-mediated gene transfer to the mammalian lung, Hum Gene Ther, 8 (1997) 689-707. [25] J. Yang, L. Zhang, C. Yu, X.-F. Yang, H. Wang, Monocyte and macrophage differentiation: circulation inflammatory monocyte as biomarker for inflammatory diseases, Biomark Res, 2 (2014) 1-1. [26] B.D. Freimark, H.P. Blezinger, V.J. Florack, J.L. Nordstrom, S.D. Long, D.S. Deshpande, S. Nochumson, K.L. Petrak, Cationic lipids enhance cytokine and cell influx levels in the lung following administration of plasmid: cationic lipid complexes, J Immunol, 160 (1998) 4580- 4586. [27] C.L. Hardy, J.S. LeMasurier, R. Mohamud, J. Yao, S.D. Xiang, J.M. Rolland, R.E. O’Hehir, M. Plebanski, Differential uptake of nanoparticles and microparticles by pulmonary APC subsets induces discrete immunological imprints, J Immunol, 191 (2013) 5278-5290. [28] H. Nakano, M.E. Free, G.S. Whitehead, S. Maruoka, R.H. Wilson, K. Nakano, D.N. Cook, Pulmonary CD103(+) dendritic cells prime Th2 responses to inhaled allergens, Mucosal Immunol, 5 (2012) 53-65. [29] V. Manolova, A. Flace, M. Bauer, K. Schwarz, P. Saudan, M.F. Bachmann, Nanoparticles target distinct dendritic cell populations according to their size, Eur J Immunol, 38 (2008) 1404- 1413.

276

[30] R.K. Wolff, Toxicology studies for inhaled and nasal delivery, Mol Pharm, 12 (2015) 2688-2696. [31] C. Duret, N. Wauthoz, R. Merlos, J. Goole, C. Maris, I. Roland, T. Sebti, F. Vanderbist, K. Amighi, In vitro and in vivo evaluation of a dry powder endotracheal insufflator device for use in dose-dependent preclinical studies in mice, Eur J Pharm Biopharm, 81 (2012) 627-634. [32] L. Jin, Q. Zhou, H.-K. Chan, I.C. Larson, M.W. Pennington, R.A.V. Morales, B.J. Boyd, R.S. Norton, J.A. Nicolazzo, Pulmonary delivery of the Kv1.3-blocking peptide HsTX1[R14A] for the treatment of autoimmune diseases, J Pharm Sci, 105 (2016) 650-656. [33] PennCenturyTM, Dry powder insufflator™ for mouse. http://penncentury.com/products/dry-powder-devices/dry-powder-insufflator-dp-4m/model- dp-4m/, 2016 [34] M. Hoppentocht, C. Hoste, P. Hagedoorn, H.W. Frijlink, A.H. de Boer, In vitro evaluation of the DP-4M PennCentury™ insufflator, Eur J Pharm Biopharm, 88 (2014) 153-159. [35] L. Garcia-Contreras, D. Sarubbi, E. Flanders, D. O'Toole, J. Smart, C. Newcomer, H.A. J., Immediate and short-term cellular and biochemical responses to pulmonary single-dose studies of insulin and H-MAP, Pharm Res, 18 (2001) 1685-1693.

277

CHAPTER 7

CONCLUSION, FUTURE DIRECTIONS AND

FINAL REMARKS

278

CONCLUSION

The thesis is comprised of a series of studies that characterise the potential of inhaled therapy contribute to the treatment of TB. Novel inhalable dry powder formulations containing efflux pump inhibitors (verapamil or thioridazine) and rifapentine were successfully formulated

(Chapter 2 and 3). These powders demonstrated excellent aerosol performance, enhanced in vitro activity against intracellular growth of M. tuberculosis compared to single drug treatments, and a promising in vitro cytotoxicity profile. On the other hand, rifapentine-PLGA particles regardless of monomer and MW showed good aerosol performance and higher affinity for macrophage uptake (Chapter 4). These polymeric particles were also relatively non-toxic to lung-associated cell lines, including macrophages and bronchial epithelial cells. These chapters have established that the rifapentine-containing inhalable dry powders generated as either pure drug or in polymeric formulations are suitable for further in vivo evaluation.

For an effective inhaled anti-TB regimen, patients may need to inhale a large dose of antibiotics repeatedly. The modified Aerolizer® (1pin_HR_00) is a low-cost generic device that can efficiently aerosolise antibiotic powders with various physiochemical properties, including crystalline rifapentine up to a maximum dose of 100 mg in a single capsule (Chapter 5). This device has the opportunity to ease the administration of large doses of medication and this may enhance patient adherence to the inhaled therapy.

Finally, the potential clinical risk associated with pulmonary drug delivery of high dose powder rifapentine was also addressed (Chapter 6). It was found that intratracheal delivery of rifapentine (20 mg/kg) triggered a transient neutrophil influx in the lungs which resolved by 7 days post-dosing. These findings suggest that the inhaled rifapentine dry powder is more

279

suitable for periodic dosing (i.e. weekly) and as an adjunct to the standard oral anti-TB drug regimen.

280

FUTURE DIRECTIONS

The results from this thesis provide a strong foundation for the following ongoing or planned studies:

1. Single dose and multi-dose murine pharmacokinetic studies using the dry powder rifapentine alone, verapamil-rifapentine and thioridazine-rifapentine are ongoing. This will be followed by TB therapeutic studies where the mice will be infected with M. tuberculosis strain and treated with the drug administered via inhalation route. In parallel, the pulmonary toxicity profile of the drugs will be evaluated. The findings from this series of in vivo studies are important for determining a rational dosage regimen for TB patients.

2. The modified Aerolizer® device will be tested for its potential to deliver the anti-TB drugs in combination. If necessary, the device will undergo further modifications to enhance its performance, for example by increasing the capsule hole size to improve the powder emptying. This study will allow the device to be optimised to deliver antibiotics in combination as will be required for multi-drug therapy of TB. The device will also be tested for human lung deposition studies.

281

FINAL REMARKS

Pulmonary treatment using powder aerosols is a relatively new approach for the treatment of

TB. Although formulation development and preclinical evaluation for various anti-tubercular drugs are in their advanced stage, human clinical trials continue to lag behind. This could be due to lack of information on pulmonary physiology of TB patients and safety concern with particle deposition in the lungs. Therefore, learning from the application of inhalation treatment of other respiratory diseases could contribute to advancing this therapeutic strategy for TB. This is a rational and rapid approach to progress inhaled TB regimen from theory to clinical reality.

Furthermore, if the anti-TB drugs manufactured into inhalable forms are expensive comparable to oral therapy, this will reduce the accessibility of the therapy to countries with the highest burden of TB cases which are typically low-income countries. If this hurdle can be mitigated satisfactorily, inhaled anti-TB therapy has potential to significantly contribute to the global elimination of TB. Although an inhaled drug regimen was not included in the recent WHO’s agenda of “END TB Strategy 2015” to eliminate TB by 2030, a global mutual partnership in the future will significantly boost the progress of the inhaled therapy for TB.

282

APPENDIX

PUBLICATIONS DURING CANDIDACY,

CONFERENCE PROCEEDINGS,

SUPPLEMENTARY DATA, AND

CO-AUTHOR DECLARATION

283

APPENDIX 1. PUBLICATIONS DURING CANDIDACY

1. Parumasivam T, Chang RYK, Abdelghany S, Ye TT, Britton WJ, Chan H-K. Dry

powder inhalable formulations for anti-tubercular therapy. Adv Drug Deliv Rev.

2016;102:83-101.

2. Parumasivam T, Chan JGY, Pang A, Quan DH, Triccas JA, Britton WJ, Chan H-K. In

vitro evaluation of inhalable verapamil-rifapentine particles for tuberculosis therapy.

Mol Pharm. 2016;13:979-989.

3. Parumasivam T, Leung SSY, Quan DH, Triccas JA, Britton WJ, Chan H-K.

Rifapentine-loaded PLGA microparticles for tuberculosis inhaled therapy: preparation

and in vitro aerosol characterization. Eur J Pharm Sci. 2016;88:1-11.

4. Parumasivam T, Chan JGY, Pang A, Quan DH, Triccas JA, Britton WJ, Chan H-K. In

vitro evaluation of novel inhalable dry powders consisting of thioridazine and

rifapentine for rapid tuberculosis treatment. Eur J Pharm Biopharm. 2016;107:205-214.

5. Parumasivam T, Leung SSY, Tang P, Mauro C, Britton W, Chan H-K. The delivery of

high-dose dry powder antibiotics by a low-cost generic inhaler. AAPS J. 2016; 19:191-

202.

6. Parumasivam T, Ashhurst AS, Nagalingam G, Britton WJ, Chan H-K. Inhalation of

respirable particles induces pulmonary inflammation. Mol Pharm. 2016; 14: 328-335.

284

APPENDIX 2. CONFERENCE PROCEEDINGS

1. Parumasivam T, Chan JGY, Pang A, Quan DH, Triccas JA, Britton WJ, Chan H-K, In

Vitro evaluation of inhalable verapamil-rifapentine particles for rapid tuberculosis

therapy, in: 15th Asia Pacific Congress of Clinical Microbiology and Infection, Kuala

Lumpur, Malaysia, 2014.

2. Parumasivam T, Leung SSY, Quan DH, Triccas JA, Britton WJ, Chan H-K, In vitro

evaluation of rifapentine-loaded PLGA microparticles for tuberculosis inhaled therapy,

in: The 8th International Conference on Materials Science and Technology, Bangkok,

Thailand, 2014.

3. Leung SSY, Parumasivam T, Mauro C, Chan H-K, A modified Aerolizer® for delivery

of high dose dry powder therapeutics, in: Respiratory Drug Delivery (RDD®), Fajardo,

Puerto Rico, 2014.

4. Parumasivam T, Chan JGY, Quan DH, Pang A, Triccas JA, Britton WJ, Chan H-K, In

vitro evaluation of inhalable thioridazine-rifapentine particles for rapid tuberculosis

therapy, in: 1st Asia Pacific AIDS & Co-infections Conference, Hong Kong, 2016.

5. Parumasivam T, Ashhurst A, Britton WJ, Chan H-K, Assessing murine immune

response to pulmonary delivery of crystalline rifapentine particles, in: 2016 CRS

Annual Meeting, Seattle, Washington, U.S.A, 2016.

285

In Vitro Evaluation of Inhalable Verapamil-Rifapentine Particles for Rapid Tuberculosis Therapy

T. Parumasivam1*, J. G. Y. Chan1, 2, A. Pang3, D. H. Quan4, W. Britton3, J. Triccas4, and H. K. Chan1

1Advanced Drug Delivery Group, Faculty of Pharmacy, The University of Sydney, Australia 2JHL Biotech, Inc., Hsinchu County, Taiwan 3Tuberculosis Research Program, Centenary Institute, Sydney, Australia 4Infectious Diseases and Immunology, Sydney Medical School, Sydney Medical School, The University of Sydney, Australia

Background Drug efflux is a crucial resistance mechanism of Mycobacterium tuberculosis (M. tb) that prolongs treatment duration by promoting tolerance to anti-tubercular (TB) drugs and inducing intracellular bacterial survival. Compared to existing first-line anti-TB therapy, both verapamil, a FDA-approved hypertension drug that functions as an efflux inhibitor, and the first-line anti-tubercular drug, rifapentine, have separately demonstrated treatment- shortening in mouse models of TB infection. Unfortunately, anti-TB antibiotics, particularly rifamycins, have been found to induce metabolism of oral verapamil thereby reducing its bioavailability. Delivering the combination of verapamil and rifapentine directly to the lungs would be a rational approach to overcome these barriers. Additional advantages of pulmonary delivery include (1) direct targeting of infected granulomas and (2) high drug concentration at the primary site of infection. In this study, we produced an inhalable combination antibiotics formulation consisting of verapamil and crystalline rifapentine. In vitro aerosol performance and biological aspects including toxicity were reported.

Methods A combination dry powder of verapamil and crystalline rifapentine in a mass ratio of 1:3 (based upon orally recommend dosage) with 20% L-leucine was produced using co-solvent precipitation and spray drying techniques. In vitro aerosol performance of the powder was assessed using a multi-stage liquid impinger (MSLI). The minimum inhibitory concentration (MIC) of the powder on M. tb H37Ra was determined via resazurin assay. Intracellular killing efficacy of the compounds was tested on THP-1 cells, a human monocyte cell line infected with M. tb H37Ra. THP-1and human lung epithelial A549 cell lines were used to examine the in vitro cytotoxicity profile of the powders.

Results The spray dried verapamil- rifapentine combination dry powder was found suitable for aerosolization. The total fine particle fraction (FPFtotal) of verapamil and rifapentine was 77.44±1.07 and 71.49±1.97, respectively. A noteworthy synergism was observed as the combination formulation had an MIC of 0.156 ng/mL, which is significantly lower than the verapamil (250 ug/mL) and rifapentine (0.625 ng/mL) alone. Rifapentine also augmented the intracellular killing of M. tb in the presence of verapamil. All the powders were shown an acceptable half maximal inhibitory concentration (IC50) on both THP-1 and A549 cell lines.

Conclusion We have demonstrated a novel inhalable verapamil-rifapentine dry powder with good aerosol performance that shows potent activity against M. tuberculosis in in vitro models. This preliminary data warrants further in vivo studies. th The 8 International Conference on Materials Science and Technology

BM-O-11

In Vitro Evaluation of Rifapentine-Loaded PLGA Microparticles for Tuberculosis Inhaled Therapy

Thaigarajan Parumasivama, Sharon S.Y. Leunga, Diana Huynh Quanb, Jamie Triccasb, Warwick Brittonc, and Hak-Kim Chana* aAdvanced Drug Delivery Group, Faculty of Pharmacy, The University of Sydney, Australia bInfectious Diseases and Immunology, Sydney Medical School, Sydney Medical School, The University of Sydney, Australia cTuberculosis Research Program, Centenary Institute, Sydney, Australia *E-mail address corresponding author: [email protected]

Keywords: Tuberculosis, Inhalation, Poly (lactic-co-glycolic acid), Rifapentine

Tuberculosis (TB) regiment requires relatively long term oral administration of first-line anti-TB drugs for curing and preventing emergence of drug-resistance. Recently, inhaled administration of drugs encapsulated into micro polymeric particles to the lungs has attracted significant attention, as it offers several advantages compared with local and systemic delivery. These include delivering high drug concentrations to lung tissues, avoiding hepatic first-pass metabolism, prolonging drug residence in the lungs which potentially reduces dosing frequency, as well as targeting alveolar macrophage that harbours tubercle cells proliferation. Poly (lactic-co-glycolic acid) (PLGA) is amongst the most studied microparticulate systems because it (1) is a FDA approved polymer for medical, (2) allows sustained release of the encapsulated drugs to minimize the need for multiple dosing, and (3) is readily phagocyted by alveolar macrophage to enhance macrophage intracellular drug concentration. Nonetheless, limited data are available in the effects of the physiochemical properties of PLGA, such as the monomer (lactide : glycide) ratio and molecular weight, on the aerosol performance. The present study aims to address this knowledge gap. PLGA with a monomer ratio of 50:50, 75:25 and 85:15 and molecular weight ranged 24-240k was studied using rifapentine (RPT) as a model drug. PLGA and RPT were dissolved in acetonitrile at concentrations of 20 mg/ml and 2 mg/ml, respectively. The mixture was spray dried using a B-290 Mini spray-dryer in closed-loop with nitrogen as the drying gas. The particle morphology and size distribution were examined using scanning electron microscopy and laser diffraction respectively. The in vitro aerosol performance of the produced powders was assessed using an OsmohalerTM coupled to a next generation impactor (NGI) running at an air flow of 100L/min for 2.4 sec. Macrophage uptake and cytotoxicity profile of these particles were also examined using THP-1 (human monocytic leukemia cell line) and A549 (adenocarcinomic human alveolar basal epithelial cells). All produced particles were spherical with smooth surface and a volume median diameter (D50) was around 2μm (with a span around 2) which is a suitable size for respiratory delivery. The fine particle fraction (FPFtotal) reached above 55% for all formulations. Apparently, no significant difference was observed in FPFtotal between all formulations indicating PLGA ratio is not a crucial parameter in PLGA inhalation. Nonetheless, the distribution of drug was found to be a function of monomer ratio, with more drug incorporated on the external surface for particles with higher lactide content. Also, a higher rate of THP-1 phagocytosis was observed for PLGA 85:15 followed by 75:25 and 50:50. Furthermore, cell toxicity analysis indicated PLGA itself does not have any effect on cell viability. Though the monomer ratio and molecular weight did not show any effect in the aerosol performance of particles, the distribution of drugs within the particles was found to be affected by these parameters. This finding may have indirect influence on the controlled release profile of drugs after they are inhaled into the lungs. Respiratory Drug Delivery 2014 – Leung et al.

A Modified Aerolizer® for Delivery of High Dose Dry Powder Therapeutics

Sharon S. Y. Leung,1 Thaigarajan Parumasivam,1 Citterio Mauro,2 and Hak-Kim Chan1

1Faculty of Pharmacy, The University of Sydney, Sydney, Australia 2Plastiape S.p.A., Osnago, Italy

KEYWORDS: colistin, rifapentine, antibiotics, dry powder inhaler

INTRODUCTION

Dry powder inhalers (DPIs) are an attractive alternative to nebulizers for high dose inhalation medications (≥ 25 mg) because they can be cheaper, more convenient to use, and may reduce administration time. The potential of using DPIs to deliver high dose therapeutics to the lungs has been recognized [1-4], and may be particularly beneficial for antibiotics [5, 6] and vaccines [7]. This work demonstrates the capability of a modified Aerolizer® to aerosolize a wide range of high dose carrier-free micronized antibiotics, including colistin for treatment of gram-negative bacterial infections, and rifapentine for treatment of tuberculosis.

MATERIALS AND METHODS

An Aerolizer was modified to accommodate a hydroxypropyl methylcellulose (HPMC) size 0 capsule (Capsugel, Peapack, NJ) and the original piercing pins were changed from four small pins (0.65 mm in diameter) to a single, centrally located large pin (1.19 mm in diameter) on each side (Figure 1). In addition, the air inlet size was reduced to 1/3 of the original design based on the area of the exterior opening to increase the air velocity and turbulence inside the swirl chamber to promote particle deagglomeration. The prototype of the modified inhaler was manufactured by Plastiape S.p.A. (Osnago, Italy). The performance of the modified Aerolizer was evaluated using colistin and rifapentine spray dried powders and a multi-stage liquid impinger (MSLI, Copley Scientific Ltd, Nottingham, UK). The colistin and rifapentine powders were prepared as reported in [8] and [9], respectively. They were of two distinct morphologies: near-spherical with corrugated surface texture (colistin) and needle-shaped (rifapentine) (Figure 2). The geometric particle size of colistin powders ranged

717 718 A Modified Aerolizer® for Delivery of High Dose Dry Powder Therapeutics – Leung et al.

from 0.5 to 2 µm and the width of the rifapentine crystals was of the same magnitude. The loaded doses for colistin and rifapentine were 50 mg and 30 mg, respectively, each representing an approximately 80% capsule fill by volume. The airflow conditions of the MSLI were 60 L/min for 4 s (4 L) corresponding to a 4.0 kPa pressure drop across the inhaler. The cut-off diameters for MSLI stages 1 to 4 are 13.0, 6.8, 3.1, and 1.7 µm, respectively. Drug was recovered from the MSLI and quantified using high performance liquid chromatography using methods Figure 1. An image of the modified Aerolizer. described in [8, 9].

Figure 2. SEM images of A) colistin and B) rifapentine.

RESULTS

For both powders, approximately 14% to 18% of powder was retained in the capsule or inhaler device after aerosolization (Figure 3). The high emitted dose indicates the modified Aerolizer is capable of aerosolizing drugs of different morphologies in the range of 30 to 50 mg. Aerosolized colistin and rifapentine powders showed emitted fine particle fractions (FPFemitted) of 65% and 80%, respectively. These results indicate that deagglomeration of the spray dried antibiotic powders occurred before they exited the device.

CONCLUSIONS

The modified Aerolizer, which can accommodate a size 0 capsule, was found to be able to disperse high dose drugs (30 to 50 mg). These preliminary data demonstrate the potential of the modified Aerolizer to deliver a wide range of dry powder therapeutics. Respiratory Drug Delivery 2014 – Leung et al. 719

Figure 3. (a) Deposition profile and (b) aerosolization performance of colistin and rifapentine powders dispersed using a modified Aerolizer.

ACKNOWLEDGEMENTS

The authors would like to thank The Australian Centre for Microscopy and Microanalysis, The University of Sydney. This research was supported under Australian Research Council's Discovery Projects funding scheme (project numbers DP120102778 & 110105161).

REFERENCES

1. Young, PM, Crapper, J, Philips, G, Sharma, K, Chan, H-K, Traini, D: Overcoming dose limitations using the Orbital® multi-breath dry powder inhaler, Journal Aerosol Med and Pulm Drug Deliv. In press. doi:10.1089/jamp.2013.1080.

2. Vandevanter, DR, Geller, DE: Tobramycin administered by the TOBI® Podhaler® for persons with cystic fibrosis: A review,Med Devices (Auckl), 2011, 4: 179-88.

3. de Boer, AH, Hagedoorn, P, Westerman, EM, Le Brun, PPH, Heijerman, HGM, Frijlink, HW: Design and in vitro performance testing of multiple air classifier technology in a new disposable inhaler concept (Twincer®) for high powder doses, Eur J Pharm Sci 2006, 28(3): 171-78.

4. Young, PM, Thompson, J, Woodcock, D, Aydin, M and Price, R: The development of a novel high-dose pressurized aerosol dry-powder device (PADD) for the delivery of pumactant for inhalation therapy, J Aerosol Med 2004, 17(2): 123-28.

5. Westerman, EM, De Boer, AH, Le Brun, PPH, Touw, DJ, Roldaan, AC, Frijlink, HW, Heijerman, HGM: Dry powder inhalation of colistin in cystic fibrosis patients: A single dose pilot study, J Cyst Fibros 2007, 6(4): 284-92.

6. Newhouse, MT, Hirst, PH, Duddu, SP, Walter, YH, Tarara, TE, Clark, AR, Weers, JG: Inhalation of a dry powder tobramycin PulmoSphere formulation in healthy volunteers, Chest 2003, 124(1): 360-66.

7. LiCalsi, C, Christensen, T, Bennett, JV, Phillips, E, Witham, C: Dry powder inhalation as a potential delivery method for vaccines, Vaccine 1999, 17(13-14): 1796-803.

8. Zhou, Q, Morton, DAV, Yu, HH, Jacob, J, Wang, J, Li, J, Chan, H-K: Colistin powders with high aerosolisation efficiency for respiratory infection: Preparation andin vitro evaluation, J Pharm Sci 2013, 102(10): 3736-47.

9. Chan, JGY, Colin, CD, Ong, HX, Chan, JCY, Tyne, AS, Chan, H-K, Britton, WJ, Young, PM, Traini, D: A novel inhalable form of rifapentine, J Pharm Sci. doi: 10.1002/jps23911.

" ! ! # # % " " +,- # '( . " " # # / " #$ % & 01* 2 & ' *45( " # 8( 6-7 8( 6-7 # % * "" " # # 68+ " # 8( 6-7 68+ 9:;5 #

" # 0 ( )&& * +(,(-( $ * ( )' * (.( / * +( 0 * 1( # * .(2( $ % " # # " " 3 4% * 5 )&* 4% * " " 6 +.7 # (* . $* * <8< # . = : 3 " %

6 $ * " # 7?; / '+ . 7?- / '(+$ )'&* 4% * 6 0 3 4% * *45( " % 4% 8 % 4* 4% * # +': 3 )1 :(( 3 )1 8( 6-7 8( 6-7 $ 6# $ # *4 " (((?': 3 )1 (((: 5( 3 )1 8( 6-7 8( 8 & 6-7 $ $ # " ! " #$ A $ # " $ " 8( " $ % 6-7 # # 9 # # # " " ! # " &" $ *4 " -+': 3 )1 "" # % :( 68+ 9:;5 #

"" " '(( ) "$ " B " % %" # # # # " # "" # # " # #

" " * $ "

# " % "

" " " "

" " $

! Assessing Murine Immune Response to Pulmonary Delivery of Crystalline Rifapentine Particles

T. Parumasivam1, A.S. Ashhurst2,3, H. K. Chan1, W.J. Britton2,3 1Advanced Drug Delivery Group, Faculty of Pharmacy, The University of Sydney, Australia 2Infectious Diseases and Immunology, Sydney Medical School, The University of Sydney, Australia 3Tuberculosis Research Program, Centenary Institute, Sydney, Australia [email protected]

Purpose: We have shown that inhalable rifapentine could significantly enhance and prolong the drug concentration in the lungs which has the potential to shorten the treatment duration of tuberculosis (TB) (1). The present in vivo study characterizes the inflammatory response caused by pulmonary deposition of rifapentine particles.

Methods: 6 – 8-week old female BALB/c mice were anesthetized with xylazine/ketamine and rifapentine powder was delivered by intratracheal intubation and insufflation at a dose of 20 mg/kg. Mice were euthanized at 12, 24 hours and 7 days post treatment, and bronchoalveolar lavage (BAL) and lungs were collected. The control was untreated mice (Nil). The left lung and BAL were processed for total cell counts and immune cell characterization via flow cytometry. The right lung was fixed by inflation and immersion in 10 % neutral buffered formalin, and stained with hematoxylin and eosin (H&E) for histopathological examination.

Results: An inflammatory response characterized by a significant influx of neutrophils was observed in the lungs at 12 and 24 hours post-exposure, and the response was resolved by day 7 (Figure 1a). The BAL showed a significant recruitment of macrophages at 24 hours (Figure 1b). The histopathological analysis also supports these observations, a pulmonary vascular congestion with prominent macrophage recruitment was observed at 12 and 24 hours, and returned to normal over 7 days.

Figure 1. (a) Neutrophil count in the lung and (b) macrophage count in the BAL. Differences between treatments were evaluated by ANOVA with Bonferroni post hoc comparison (*p < 0.05, **p < 0.01).

Conclusion: Pulmonary delivery of rifapentine caused a neutrophil-associated inflammatory response in the lungs and resolved over 7 days, limiting the pulmonary delivery of rifapentine to once a week at a dose of 20 mg/kg. Efficacy of weekly dosing inhalable rifapentine in a murine TB model is planned.

References: 1. Chan JGY, Tyne AS, Pang A, McLachlan AJ, Perera V, Chan JCY, Britton WJ, Chan HK, Duke CC, Young PM, Traini D. Murine pharmacokinetics of rifapentine delivered as an inhalable dry powder. Int J Antimicrob Agents. 2014;43(3):319-323.

APPENDIX 3. SUPPLEMENTARY DATA FROM CHAPTER 4

Exothermic → Exothermic SP(1)50:50_24k SP(1)50:50_38k SP(1)50:50_54k SP(1)75:25_76k SP(1)85:15_50k SP(1)85:15_190k 22 52 82 112 142 172 202 232 262 292 Temperature (°C)

Exothermic → Exothermic EV85:15_50k EV75:15_76k EV50:50_24k SP(2)85:15_50k SP(2)75:25_76k SP(2)50:50_24k 22 52 82 112 142 172 202 232 262 292 Temperature (°C)

Figure 1. DSC thermograms showing glass transition temperature (Tg) (°C) of PLGA- rifapentine particles. The samples were heated 30-300°C under a dynamic flow of nitrogen gas.

294

Figure 2. Agglomeration of SP(1)85:15_190k exposed to dissolution media for 24 h, viewed under scanning electron microscope.

295

a 120 100

80 SP(1)50:50_24k 60 SP(1)50:50_38k SP(1)50:50_54k 40 SP(1)75:25_76k Cell Cell viability (%) SP(1)85:15_50k SP(1)85:15_190k 20 Rifapentine

0 0 0.5 1 1.5 Particle concentration (µg/mL)

b 120 100

SP(1)50:50_24k 60 SP(1)50:50_38k SP(1)50:50_54k

40 SP(1)75:25_76k Cell Cell viability (%) SP(1)85:15_50k 20 SP(1)85:15_190k Rifapentine 0 0 0.5 1 1.5 Particle concentration (µg/mL)

Figure 3. (a) THP-1 and (b) A549 cell viability (MTT assay) after 4 days of exposure to rifapentine loaded various PLGA particles (mean ± SD, n=3). None of the formulations led to a reduction of cell viability below 50 %, indicates that the PLGA particles non-toxic to respiratory associated cell lines.

296

APPENDIX 4. SUPPLEMENTARY DATA FROM CHAPTER 5

Table 1. Laser diffraction analysis showing volume equivalent diameters of the particles.

D10 (µm) D50 (µm) D90 (µm)

Colistin 0.77 ± 0.03 1.57 ± 0.04 3.06 ± 0.11

Rifapentine 0.61 ± 0.13 1.41 ± 0.13 3.07 ± 0.30

Tobramycin 0.55 ± 0.07 1.31 ± 0.11 2.73 ± 0.10

297

Table 2. Performance of 1pin_HR_00 dispersing colistin at various dispersion conditions.

Flow Time Inhaled Amount of powder FPFRecovered Comments (L/min) (sec) air emitted in the first (%) volume actuation (L) (mg) 30 2.0 1.0 ~ 11.1 19.0 ± 0.1 A low FPF 30 3.0 1.5 ~ 13.0 n/d A low FPF is expected 30 4.0 2.0 ~ 13.0 n/d A low FPF is expected 45 1.4 1.0 ~ 20.0 41.8 ± 1.4 A low FPF 45 2.0 1.5 ~ 21.1 46.9 ± 3.3 An acceptable FPF and emptying dose 45 2.7 2.0 ~ 32.0 51.0 ± 1.5 An acceptable FPF and emptying dose 60 1.0 1.0 ~ 62.1 n/d Emptying a relatively large amount 60 1.5 1.5 n/d n/d Emptying a relatively large amount is expected 60 2.0 2.0 n/d n/d Emptying a relatively large amount is expected n/d : not determined

298

Table 3. Performance of 1pin_HR_00 dispersing rifapentine at various dispersion conditions.

Flow Time Inhaled Amount of powder FPFRecovered Comments

(L/min) (sec) air emitted in the first (%)

volume actuation

(L) (mg)

30 2.0 1.0 n/d n/d A low FPF is

expected

30 3.0 1.5 n/d n/d A low FPF is

expected

30 4.0 2.0 ~ 1.8 3.9 ± 0.1 A low FPF

45 1.4 1.0 n/d n/d A low FPF is

expected

45 2.0 1.5 ~ 5.0 15.9 ± 1.0 A low FPF

45 2.7 2.0 ~ 7.9 14.3 ± 1.0 A low FPF

60 1.0 1.0 ~ 8.3 n/d A low FPF is

expected

60 1.5 1.5 ~ 8.1 16.0 ± 1.7 A low FPF

60 2.0 2.0 ~ 10.7 58.4 ± 2.4 An acceptable

FPF and

emptying dose n/d : not determined

299

Tobramycin

Rifapentine Intensity

Colistin

10 15 20 25 30 35 40 2 θ/Degree

Figure 1. X-ray powder diffraction patterns of the spray-dried antibiotics.

Tobramycin

Rifapentine

Exothermic → Exothermic Colistin

30 60 90 120 150 180 210 240 270 300 Temperature (°C)

Figure 2. DSC thermograms of the spray dried-drug powders.

300

60 Colistin Rifapentine 50 Tobramycin

20 Change in mass (%) mass in Change

0 0 10 20 30 40 50 60 70 80 90 Relative humidity (%)

Figure 3. Dynamic vapour sorption behaviour of the spray-dried drug powders.

301

APPENDIX 5. SUPPLEMENTARY DATA FROM CHAPTER 6

A 4 0 *** ** B 4 5 *

) 4 0

3 0

)

(

s l

( 3 5

h l

2 0 l

o r

c 3 0

u B e 1 0 N 2 5

0 2 0 N il 1 2 h 2 4 h 7 d N il 1 2 h 2 4 h 7 d

C 0 .5 D 1 .0

0 .4 0 .8

)

(

)

0 .3 o 0 .6

(

C m

0 .2 / 0 .4

a M 0 .1 0 .2

0 .0 0 .0 N il 1 2 h 2 4 h 7 d N il 1 2 h 2 4 h 7 d

E 1 5 F 0 .3 )

(

)

s 1 0 0 .2

(

p A

r 5 0 .1

a m

0 0 .0 N il 1 2 h 2 4 h 7 d N il 1 2 h 2 4 h 7 d

G * 4 0

3 0

)

(

l 2 0

- T 1 0

0 N il 1 2 h 2 4 h 7 d

302

Figure 1. Immune cell subsets in the lungs at 12 h, 24 h and 7 day post-treatment as a percentage of CD45.2+ cells. The control mice were untreated (Nil). (A) Neutrophils (Ly6G+

CD11b+), (B) B-cells (B220+ CD11c-), (C) dendritic cells (DC) (B220- F4/80- CD11chi), (D) combined macrophage/monocyte populations (Mac/mono) (B220- F4/80- CD11b+ CD11c-), (E) alveolar macrophages (B220- F4/80+ CD11b+ CD11c-), (F) plasmacytoid dendritic cells (pDC)

(B220+ CD11c+), and (G) T-cells (CD3+). Differences between treatments were evaluated by

ANOVA with Bonferroni post hoc comparison (*p < 0.05, **p < 0.01, ***p < 0.001).

303

**** A B 2 0 * **** 1 0 0

1 5

) %

) 8 0

(

l l s 1 0

l 6 0

o B

r 4 0

t 5

u e

N 2 0

0 0 N il 1 2 h 2 4 h 7 d N il 1 2 h 2 4 h 7 d

C 0 .2 0 D ** ** 6

0 .1 5 )

( )

( n

0 .1 0

/ D

c 2 a

0 .0 5 M

0 .0 0 0 N il 1 2 h 2 4 h 7 d N il 1 2 h 2 4 h 7 d

E **** **** F 0 .3

2 5 )

2 0

(

) 0 .2

e a

1 5 (

p l

1 0 p o

A 0 .1

c a

5 m

0 0 .0 N il 1 2 h 2 4 h 7 d N il 1 2 h 2 4 h 7 d

G 8 0 **** ****

6 0

)

(

l 4 0

- T 2 0

0 N il 1 2 h 2 4 h 7 d

304

Figure 2. Immune cell subsets in the bronchoalveolar lavage (BAL) at 12 h, 24 h and 7 day post-treatment as a percentage of CD45.2+ cells. The control mice were untreated (Nil). (A)

Neutrophils (Ly6G+ CD11b+), (B) B-cells (B220+ CD11c-), (C) dendritic cells (DC) (B220-

F4/80- CD11chi), (D) combined macrophage/monocyte populations (Mac/mono) (B220- F4/80-

CD11b+ CD11c-), (E) alveolar macrophages (B220- F4/80+ CD11b+ CD11c-), (F) plasmacytoid dendritic cells (pDC) (B220+ CD11c+), and (G) T-cells (CD3+). Differences between treatments were evaluated by ANOVA with Bonferroni post hoc comparison (*p < 0.05, **p

< 0.01, ***p < 0.001).

305

APPENDIX 6. CO-AUTHOR DECLARATION

Declaration from the co-authors of publications included within this thesis as chapters are listed in the following pages.

1. Professor Hak-Kim Chan

2. Professor Warwick Britton

3. Associate Professor Jamie Triccas

4. Associate Professor Sharif Abdelghany

5. Dr. John Gar Yan Chan

6. Dr. Sharon Shui Yee Leung

7. Dr. Yoon Kyung (Rachel) Chang

8. Dr. Patricia Tang

9. Dr. Tian Tian Ye

10. Ms. Gayathri Nagalingam

11. Ms. Angel Pang

12. Ms. Diana Huynh Quan

13. Mrs. Anneliese Ashhurst

14. Mr. Mauro Citterio

306

To whom it may concern,

I, Associate Professor Jamie Triccas, declare that Thaigarajan Parumasivam was the primary researcher within the papers submitted in this thesis. All experimental work and manuscript presentation were conducted by Thaigarajan, except as specified in the following publications wherein I acted in an advisory capacity.

1. T. Parumasivam, J.G.Y. Chan, A. Pang, D.H. Quan, J.A. Triccas, W.J. Britton, H.K.

Chan, In vitro evaluation of inhalable verapamil-rifapentine particles for tuberculosis

therapy, Mol Pharm, 13 (2016) 979-989.

2. T. Parumasivam, J.G.Y. Chan, A. Pang, D.H. Quan, J.A. Triccas, W.J. Britton, H.K.

Chan, In vitro evaluation of novel inhalable dry powders consisting of thioridazine

and rifapentine for rapid tuberculosis treatment, Eur J Pharm Biopharm, 107 (2016)

205-214.

3. T. Parumasivam, S.S.Y. Leung, D.H. Quan, J.A. Triccas, W.J. Britton, H.K. Chan,

Rifapentine-loaded PLGA microparticles for tuberculosis inhaled therapy: preparation

and in vitro aerosol characterization, Eur J Pharm Sci, 88 (2016) 1-11.

Signed: Date: 01 September 2016

To whom it may concern,

I, Dr. John Gar Yan Chan, declare that Thaigarajan Parumasivam was the primary researcher within the papers submitted in this thesis. All experimental work and manuscript presentation were conducted by Thaigarajan, except as specified in the following publications wherein I acted in an advisory capacity.

1. T. Parumasivam, J.G.Y. Chan, A. Pang, D.H. Quan, J.A. Triccas, W.J. Britton, H.K.

Chan, In vitro evaluation of inhalable verapamil-rifapentine particles for tuberculosis

therapy, Mol Pharm, 13 (2016) 979-989.

2. T. Parumasivam, J.G.Y. Chan, A. Pang, D.H. Quan, J.A. Triccas, W.J. Britton, H.K.

Chan, In vitro evaluation of novel inhalable dry powders consisting of thioridazine

and rifapentine for rapid tuberculosis treatment, Eur J Pharm Biopharm, 107 (2016)

205-214.

Signed: John Gar Yan Chan Date: 01 September 2016

To whom it may concern,

I, Dr. Yoon Kyung (Rachel) Chang, declare that Thaigarajan Parumasivam was the primary researcher within the papers submitted in this thesis. All experimental work and manuscript presentation were conducted by Thaigarajan, except as specified in the following publications wherein I contributed to the writing of a section in the manuscript.

1. T. Parumasivam, R.Y.K. Chang, S. Abdelghany, T.T. Ye, W.J. Britton, H.K. Chan,

Dry powder inhalable formulations for anti-tubercular therapy, Adv Drug Deliv Rev,

102 (2016) 83-101.

Signed: Date: 01 September 2016

To whom it may concern,

I, Dr. Tian Tian Ye, declare that Thaigarajan Parumasivam was the primary researcher within the papers submitted in this thesis. All experimental work and manuscript presentation were conducted by Thaigarajan, except as specified in the following publications wherein I contributed to the writing of a section in the manuscript.

1. T. Parumasivam, R.Y.K. Chang, S. Abdelghany, T.T. Ye, W.J. Britton, H.K. Chan,

Dry powder inhalable formulations for anti-tubercular therapy, Adv Drug Deliv Rev,

102 (2016) 83-101.

Signed: Date: 01 September 2016

“The man who graduates today and stops learning tomorrow is uneducated the day after”

- Newton D. Baker -

323