Journal of Controlled Release 323 (2020) 412–420

Contents lists available at ScienceDirect

Journal of Controlled Release

journal homepage: www.elsevier.com/locate/jconrel

Excipient-free pulmonary insulin dry powder: Pharmacokinetic and T

pharmacodynamics profiles in rats
Eride Quartaa,f,1, Veronica Chiericib,1, Lisa Flamminia, Massimiliano Tognolinia,
Elisabetta Barocellia, Anna Maria Cantonic, Gabriela Dujovnyd, Susana Ecenarro Probstd,
Fabio Sonvicoa, Gaia Colomboe, Alessandra Rossia, Ruggero Bettinia, Paolo Colomboa,f,
Francesca Buttinia,
⁎

a
Food and Drug Department, University of Parma, Parco Area delle Scienze 27/A, 43124 Parma, Italy
b
Interdepartmental Center for Innovation in Health Products, Biopharmanet_TEC, University of Parma, Parco Area delle Scienze 27/A, 43124 Parma, Italy
c
Department of Veterinary Health Sciences, University of Parma, Strada del Taglio 8, 43126 Parma, Italy
d
Qualicaps, Avenida Monte Valdelatas 4, 28108 Alcobendas, Madrid, Spain
e
Department of Life Sciences and Biotechnology, Via Fossato di Mortara 17/19, 44121 Ferrara, Italy
f
Paolo Colombo and Eride Quarta. f: PlumeStars Srl, Strada Inzani 1, 43125 Parma, Italy

ARTICLE INFO ABSTRACT

Keywords: A novel pure insulin spray-dried powder for DPI product (Ins_SD) was studied with respect to physico-chemical
Insulin stability, in vitro respirability, bioavailability, activity and tolerability.
Excipient-free DPI Ins_SD powder exhibited a very high in vitro respirability, independently of the DPI product preparation
Dry powder inhaler (manual or semi-automatic). Physico-chemical characteristics of Ins_SD powder remained within the pharma-
Stability
copoeia limits during 6 months of storage at room temperature.
Pharmacokinetics
Pharmacodynamics
PK/PD profiles were measured in rats that received the pulmonary powders by intratracheal insufflation and
Safety compared with Afrezza inhalation insulin. Due to the low drug powder mass to deliver, both insulin powders
were diluted with mannitol. Insulin from Ins_SD was promptly absorbed (tmax 15 min and Cmaxx4.9 ± 1.5 mU/
ml). Afrezza had a slower absorption (tmax 30 min and Cmax of 1.8 ± 0.37 mU/ml). After glucose injection,
Ins_SD determined a rapid reduction of glucose level, similar to Afrezza. As reference, insulin subcutaneous
injection showed a long-lasting hypoglycemic effect due to the slow absorption that prolonged insulin plasma
level.
In summary, Ins_SD product is suitable for post-prandial glucose control, providing a convenient and com-
pliant product, in particular in the event of using a disposable device.
Albeit the product has to be stored in fridge, its stability at room temperature allows the diabetic individual to
carry the daily dose in normal conditions.

1. Introduction fear of the administration by injection, thus increasing the patient's

convenience of the therapy [2]. Treatment with insulin by oral in-
Approximately 6 million people in the U.S. require insulin therapy halation addressed these concerns and empowered patients to obtain
for glycemic control [1]. Insulin is mainly administered subcutaneously appropriate glycemic control. It has been a long time since pulmonary
or intravenously. Some patients are disturbed by insulin administration insulin both as liquid and dry powder was investigated [3]: Aerami
because of adverse-events (e.g., hypoglycemia, weight gain) and fear of therapeutics has completed a Phase 2 clinical trial of a liquid for-
injections. Inhalation delivery of insulin avoids the invasiveness and mulation delivered as a soft mist (web ref 1). Pfizer company had

Abbreviations: AUC, Area Under the Curve; BALF, Bronchoalveolar lavage fluid; DPI, Dry Powder Inhaler; FDA, Food and Drug Administration; FDKP, Fumaryl-
diketopiperazine; FPD, Fine Particle Dose; FPF, Fine Particle Fraction; HMWP, high molecular weight protein; IT, Intratracheal; Mann, Mannitol; MMAD, mass
median aerodynamic diameter; ORP, other related proteins; PD, Pharmacodynamic; PK, Pharmacokinetic; RBA, Relative Bioavailability; SC, Subcutaneous
⁎
Corresponding author at: Food and Drug Department, University of Parma, Parco Area delle Scienze, 27/A, 43124 Parma, Italy.
E-mail address: francesca.buttini@unipr.it (F. Buttini).
1
Equal contribution to the paper by first two authors.

https://doi.org/10.1016/j.jconrel.2020.04.015
Received 27 February 2020; Received in revised form 4 April 2020; Accepted 9 April 2020
Available online 21 April 2020
0168-3659/ © 2020 Elsevier B.V. All rights reserved.
E. Quarta, et al. Journal of Controlled Release 323 (2020) 412–420

developed an inhaled insulin product that obtained the US FDA ap- 2. Materials and methods
proval in 2006 with the brand name of Exubera®. In October 2007,
Pfizer withdrew the product from the market due to unsatisfying sales, 2.1. Materials
in part due to serious side effects and risks (hypoglycemia, cough,
pharyngitis, and rhinitis) [4] and in part attributed to the device size Human recombinant insulin (CAS Number 11061–68-0) was pur-
[5,6]. In July 2014, Afrezza® (human insulin) inhalation powder chased from Sigma Aldrich (Sigma Chemical Co., St. Louis, MO, US). A
(MannKind Corp, US), was approved by FDA for type-1 and type-2 pack of Afrezza® (insulin human) Inhalation Powder containing single-
diabetes mellitus patients. It employs a small and easy-to-use dry use cartridges of 4 units and 8 units (batch number 000125P, MannKind
powder inhaler. This novel insulin reduces the glucose levels similarly Corp./Sanofi-Aventis US) was purchased. Mannitol for inhalation was
to rapid-acting insulin injection, with better control of hypoglycemia supplied by Merck KGaA (Darmstadt, DE). HPMC extra-dry capsules for
[7–9]. Upon inhalation, insulin microparticles are deposited into the use in dry powder inhalers, Quali-V®-I size #3, were provided by
deep lung where they dissolve. The absorbed peptide enters the Qualicaps (Madrid, ES) and a single dose high resistance (0.031
bloodstream producing the glucose-lowering response. The pharmaco- kPa ·(L /min) 1 ) dry powder inhaler RS01_HR was gifted by Plastiape
dynamic effect appears quickly and has a shorter duration than the (Lecco, IT). Medical gray sheets for packaging of polypropylene/alu-
subcutaneous injected products [10]. minum/polyvinyl chloride/polyvinylidene chloride (PP/AL/PVC/
To favour and control the pulmonary deposition and absorption of PVDC) were supplied by Research Pharmaceutical (Shanghai, CN) and
insulin, a number of studies have used excipients as carriers and ab- aluminum foil was purchased by Amcor Flexibles (Soliera, IT). All other
sorption promoters [11–13]. For example, Jiake He et al. (2018) stu- chemicals used were obtained from commercial suppliers and were of
died pharmacokinetic and pharmacodynamic performance of dry analytical grade.
powders of rh-insulin administered intratracheally in rats, formulated
with pig pulmonary surfactant or phospholipid hexadecanol tyloxapol
[12]. Conjugation with biocompatible substrates or polymers, with the 2.2. Preparation of insulin and mannitol spray-dried powders
intent to construct highly respirable insulin particles and resistant to
proteolytic degradation of alveolar macrophages, was as well described A human recombinant insulin powder for inhalation (Ins_SD) was
[11]. prepared by spray drying as described before [18,19]. In detail, the
Exubera insulin powder was formulated with a number of excipients micronized insulin powder was obtained starting from 1% (w/v) solu-
(calcium chloride, phospholipids, citric acid) to obtain highly respirable tions of the peptide. Human insulin raw material was dissolved in acetic
particles, named PulmoSphere™ [14]. Currently, the only oral inhala- acid 0.4 N providing a pH of 2.6, then the pH was raised to 3.6 using
tion powder formulation on the market (Afrezza, human insulin) con- 10% (v/v) ammonium hydroxide. Spray-dried insulin powder was
tains 18% w/w of insulin diluted with fumaryl-diketopiperazine (FDKP) manufactured employing a Büchi Mini Spray Dryer B-191 (Büchi La-
and polysorbate 80 [15]. These excipients were used to manufacture bortechnik, Flawil, CH). Briefly, 100 ml of insulin solution was sprayed
respirable microparticles, according to the proprietary Technosphere® using the following process parameters: inlet temperature 120 °C,
technology [16]. According to the manufacturer, FDKP has no biolo- drying air flow rate 600 l/h, aspiration 35 m3/h, solution feed rate of
gical activity and acts as vehicle for delivering insulin to the alveoli 3.5 ml/min and a nozzle diameter 0.7 mm. Under these conditions an
[15,17]. outlet temperature of 40–60 °C was measured.
The presence of excipients in the formulations for chronic pul- In order to have an accurate mass of insulin for the in vivo study by
monary administration of drugs is scrutinized by regulatory agencies. inhalation, a mannitol spray-dried microparticulate powder (Mann_SD)
The insulin lung powders like Exubera and Afrezza have to be stored was prepared to dilute the insulin powders. Mann_SD powder was spray
and kept at 2–8 °C; consequently, the product for daily therapy requires dried from a 1% (w/v) water solution of mannitol at an inlet tem-
the availability of cold storage ([1], web ref 2). perature of 160 °C. All the other process parameters were set as de-
A pure insulin pulmonary powder was invented at University of scribed for insulin powder preparation.
Parma [18]. The novel formulation of insulin for inhalation (Ins_SD)
was manufactured as spray-dried powder free of excipient. The novelty
resided in the preparation of the respirable powder by spray drying of 2.3. Filling and packaging of Quali-V®-I capsules
an aqueous acetic solution of peptide. Insulin spray dried from acidic
solution was sufficiently stable at room temperature [19] and the re- HPMC size #3 capsules (Quali-V®-I) were semi-automatically filled
frigeration over short periods of patient use could be less stringent. with 2 mg of Ins_SD powder using a Omnidose TT vacuum drum filler
The aim of this work was to investigate the novel pure insulin spray- system (Harro Höfliger, DE). A batch size of 100 capsules was prepared.
dried powder as component of DPI product, studying physico-chemical Then, the capsules were packed in a blister constituted by preformed
stability, in vitro respirability, bioavailability, activity and tolerability. foil of PP/AL/PVC/PVDC with a thickness of 350 μm and sealed with a
In particular, with the objective to develop the inhalation product, the standard 20 μm thickness aluminum foil. The sealing process was con-
effects of filling process and capsule/blister packaging were assessed. ducted at a temperature of 140 °C, for 1 s at a pressure of 1410 N/cm2.
The stability studies were conducted storing the capsules, packed in A small batch of 10 capsules was also prepared by filling the cap-
blisters, at room temperature and in refrigerated conditions for up to six sules manually with 3 mg of powder in order to assess the effect of the
months. The in vitro respirability was studied during the storage period. semi-automatic filling on the aerodynamic performance.
Covalent degradation products with high molecular weight and insulin
related proteins were searched.
Finally, the PK/PD in rats after intratracheal insufflation of Ins_SD at 2.4. Insulin stability test
a dose of 10 IU/kg was investigated in comparison with Afrezza and
subcutaneous administration of insulin solution. The plasma con- Ins_SD, dosed in HPMC capsules sealed in blisters, was stored at
centration profile of insulin and the pharmacodynamic effect, i.e., the 25 °C and 60% RH. Samples were analyzed for degradation products of
level of plasma glucose, were determined. In addition, the potential A21, other related proteins (ORP) and high molecular weight protein
inflammatory effects of the powder insufflation were studied by histo- (HMWP) at 0, 30, 90 and 180 days after powder filling. The same sta-
logical analysis of animal lung tissue. The granulocyte and monocyte bility tests were performed on blistered powders stored at 4 °C. The
populations present in the bronchoalveolar lavage fluid were de- aerodynamic performance of the Ins_SD powder was determined using
termined. NGI at 0, 30, 90 and 180 days after powder filling.

413
E. Quarta, et al. Journal of Controlled Release 323 (2020) 412–420

2.5. Insulin and insulin degradation products assay was drawn through the inhaler during each experiment.
The test was executed on both manually-filled and semi-auto-
Insulin content was determined by HPLC using the conditions re- matically-filled Ins_SD capsules. The content of three capsules of each
ported in USP42. Test and reference solutions were prepared in HCl type was discharged for each NGI test. The number of Afrezza car-
0.01 N. The stationary phase was a C18 Symmetry300™ column tridges discharged in the NGI for each test was six for the 8 units test
(250 × 4.6 mm, 5 μm, Waters Corp., MA, US). The quantification and nine for the 4 units test. According to the product leaflet, the 8 units
method was linear over the concentration range between 10 and 60 μg/ cartridge and the 4 units cartridge contain 0.7 mg and 0.35 mg of in-
ml (R2 = 0.999) and the precision was 0.55%. sulin, respectively, equivalent to 20 and 10 IU of insulin.
The content of impurities with molecular masses greater than that of The metered dose (MD) is the mass of drug quantified by HPLC,
insulin and related proteins were determined. In particular, A21 desa- calculated by summing the drug recovered from the inhaler and the
mido insulin and other related proteins (ORP) were determined by li- impactor (induction port, stages 1 to 7 and MOC). The Emitted Dose
quid chromatography using the gradient conditions according to (ED) is the amount of drug leaving the device and entering the impactor
USP42. Concerning the high molecular weight protein (HMWP) im- (induction port, stages 1 to 7 and MOC). The mass median aerodynamic
purities, size exclusion chromatography was performed using an Insulin diameter (MMAD) was determined by plotting the cumulative percen-
HMWP column (300 × 7.8 mm, Waters Corp., Milford, MA, US). The tage of mass less than the stated aerodynamic diameter for each NGI
test solution was prepared by dissolving 4 mg of insulin in 1 ml of HCl stage on a probability scale versus the aerodynamic diameter of the
0.01 N. The peak-to-valley ratio between the height of the dimer peak stage on a logarithmic scale. The Fine Particle Dose (FPD) is the mass of
and the height of the lowest point of the curve separating dimer and drug < 5 μm and the Extra-Fine Particle Dose (Extra-FPD) is the mass of
monomer peak, determined to assess the system suitability, was higher drug < 2 μm (both calculated from log-probability plot equation). The
than 2.0. Fine particle fraction (FPF) and the Extra-fine particle fraction (Extra-
FPF) are the ratio between FPD or Extra-FPD and ED in percent, re-
2.6. Particle-size distribution by laser diffraction spectively.

The particle size distribution of Ins_SD was determined using a 2.8. Particle morphology by scanning electron microscopy
Spraytec® (Malvern Instruments Ltd., Worcestershire, UK) laser dif-
fraction equipped with a 45 mm focal lens, which measures particle size Scanning electron microscopy (SEM, Zeiss AURIGA, Oberkochen,
in the range from 0.1 to 80 μm. Briefly, 10 mg of powder were dispersed DE) was operated under high vacuum conditions with an accelerating
in cyclohexane containing 0.1% (w/v) of lecithin and samples were 1.0 kV voltage, at different magnifications. Powders were deposited on
sonicated for 5 min. The results of the analysis were expressed as dv,90, adhesive black carbon tabs pre-mounted on aluminum stubs and im-
dv,50, and dv,10 i.e., cumulative undersize volume diameters at 90%, aged without undergoing any metallization process.
50% and 10% of particle population, respectively.
The particle size distribution of Ins_SD/mannitol or Afrezza/man- 2.9. Pharmacokinetic and pharmacodynamics profile of pulmonary insulin
nitol blends upon aerosolization from DP-4 insufflator was assessed in formulations
dry conditions. The DP-4 insufflator loading chamber was filled with
20 mg of blend and aerosolized with 10 ml of air in front of the laser. PK/PD profiles of Ins_SD pulmonary administration were in-
The aerosol cloud was arranged to cross the laser beam perpendicularly. vestigated in comparison to Afrezza and subcutaneous injection. The
The end of the device needle was positioned at 2.5 cm from the laser experiments were performed with the approval of the Local Ethical
beam and at 1.0 cm from the detector. The measure was carried out in Committee of University of Parma and by the Italian Ministry of Health,
triplicate. in according to the Italian (DLgs 26/2014) and the European (Directive
2010/63/EU) legislation. 36 male Wistar rats (Charles River
2.7. Aerodynamic assessment of Afrezza and Ins_SD Laboratories International, LC, Italy) weighing 280–320 g were housed
under standard conditions (22 ± 2 °C, humidity 55 ± 10%, 12 h light
The aerodynamic performance of insulin batches was assessed using and dark cycle, food and water available ad libitum).
the Next Generation Impactor (NGI; Copley Scientific, UK). The NGI The animal groups and the formulation tested are summarized in
was connected to a SCP5 vacuum pump (Copley Scientific, UK) through Table 1. The insulin dose administered to rats was 10 IU/kg. Con-
a TPK. The cups of the impactor were coated with a thin layer of sidering the animal weight, the amount of insulin powder to insufflate
ethanol containing 2% (w/v) of Tween 20 to prevent particle bounce. into the lungs was in the range of 100–110 μg. This amount was too low
Insulin was recovered from the apparatus by washing with HCl 0.01 N to be accurately weighed and loaded in the insufflator device. The
and assayed by HPLC as previously reported. strategy adopted to overcome this issue was to prepare powder blends
Due to the resistance of the DreamBoat™ inhaler of Afrezza, the in having an insulin content of 4% w/w using spray-dried mannitol
vitro respirability of this product was tested at 30 L/min air flow rate in (Mann_SD) as bulking excipient. Mannitol was selected due to its safety
order to achieve a 4 kPa pressure drop. The flow rate was set using a [20] and chemical compatibility with insulin. Ins_SD and Afrezza
DFM 2000 Flow Meter (Copley Scientific, UK). Ins_SD was aerosolized blends with spray-dried mannitol were mixed manually in a glass
using RS01_HR device and this device was activated at 65 L/min. In all mortar for ten minutes. The drug content and homogeneity were as-
the tests, the duration time was adjusted so that a volume of 4 L of air sessed by HPLC. The mannitol blends of Ins_SD and of Afrezza were

Table 1
Treatment groups for the evaluation of insulin and glycemic profile in rats after intratracheal (IT) administration (n = 9). Group 4 received
an insulin solution via subcutaneous (SC) injection.
Treatment group Route of administration Type of administration

Group 1 IT Excipient microparticles (Mann_SD)

Group 2 IT Ins_SD/Mannitol_SD powder blend (Ins_SD/mann)
Group 3 IT Afrezza/Mannitol_SD powder blend (Afrezza/mann)
Group 4 SC Insulin solution 10 IU/ml (Ins SC)
Group 5 IT 4 ml air puff

414
E. Quarta, et al. Journal of Controlled Release 323 (2020) 412–420

weighed directly in the insufflator at an insulin dose of 10 UI/kg cor- Diego, CA, US) for 1 h in the dark at 4 °C. Cells were then washed with
responding roughly to a final weight of 2–3 mg of blended powder. PBS and suspended in cell staining buffer. The viability of the cellular
Mann_SD was intratracheally administered to the control group at a suspension was determined through propidium iodide (PI) staining:
dose of 8 mg/kg (group 1). Insulin solution was administered sub- cells were incubated with PI 10 μg/ml (BioLegend™, San Diego, CA, US)
cutaneously to group 4 by dissolving Ins_SD in saline solution (0.9% for 1 min in the dark, at room temperature, and immediately subjected
NaCl w/v) with the addition of 9% (v/v) of HCl 0.01 N in order to to flow cytometry analysis. Only PI- cells were included in the analysis.
obtain a clear solution quickly. The insulin concentration in the solution Samples were analyzed using Guava easyCyte™ and InCyte™ soft-
prepared was 10 IU/ml and the final pH was 4.7. Finally, a group of ware (Merck Millipore, Darmstadt, DE). Leukocytes were defined by
animals employed as control received intratracheally a puff of 4 ml air gating on granuolocytes-monocytes (CD45+ CD3- CD43+). The results
by DP-4 insufflator™. were reported as mean ± standard error for each treatment. The sta-
Animals were fasted for 3 h prior to basal glycemic determination; tistical analysis was carried out by one-way ANOVA followed by the
then, they received 1 g/kg glucose by intraperitoneal injection (time Dunnett's post-test.
zero) and 5 min later 10 IU/kg of insulin were administered either Lungs were fixed in 10% neutral buffered formalin; after paraffin
subcutaneously (SC) or intratracheally (IT). DP-4 insufflator™ dry embedding, 5 μm thick sections were obtained with microtome (Leica)
powder inhaler (Penn-Century, Inc., Philadelphia, US) device was used and stained with Hematoxylin and Eosin (H&E). Histological slides
for IT delivery. In this case, the animals were anesthetized by an in- were examined with Nikon Eclipse E800 microscope (Nikon
traperitoneal injection of Nembutal (50 mg/kg). After falling asleep, Corporation, JP). Sections were photographed at 4×, 10×, 20× and
each rat was laid on its back at an angle of 45°. A clear view of the 40× (Nikon PLAN APO lenses) with Camera DIGITAL SIGHT DS-Fi1
trachea was provided by inserting a fibre optic laryngoscope (Welch- (Nikon Corporation, JP) and pictures were acquired with DS Camera
Allyn) into the mouth of the animal. Then, the device tip of the DP-4 Control Unit DS-L2 (Nikon Corporation, JP). Histological lesions were
insufflator attached to air-filled syringe was inserted into the trachea. graded and classified based on the extension/distribution (focal, mul-
Subsequently, the powder particles loaded in the insufflator were de- tifocal and diffuse) and severity (scant, mild, moderate, severe).
livered into the lung discharging 4 ml of air.
Blood samples (200 μl) were collected by sublingual vein at time 15, 3. Results and discussion
30, 60, 120 and 180 min after administration of glucose. The rats were
sacrificed under anaesthesia after completion of the experimental pro- The Ins_SD powder was successfully manufactured by spray drying
cedure by carbon dioxide inhalation. The blood samples were collected with a process yield of 72%. The powder exhibited a median volume
into micro-centrifuge tubes containing EDTA as anticoagulant. Plasma diameter of 2.72 μm and a dv,10 and dv,90 of 1.20 and 7.91 μm, re-
was obtained by blood samples centrifugation at 10000 rpm for 5 min. spectively. As previously reported, the thermogravimetric analysis as-
Glucose concentration was determined using a Glucose Oxidase reagent sociated with the FTIR spectrometry analysis showed that Ins_SD had a
(Instrumentation Laboratory-Werfen). The results were reported as weight loss of about 5% in the range 30–100 °C, attributable to water
mean ± standard error for each treatment. The statistical analysis was loss. No presence of acetic acid/ammonium hydroxide was detected
carried out by two-way ANOVA followed by the Bonferroni's post-test. [19]. Thus, the spray drying process removed the volatile acetic acid
For insulin determination ELISA kit (Crystal Chem 90,095), a from the Ins_SD powder. The SEM analysis showed that insulin raw
sandwich ELISA assay suitable for the determination of human insulin material (Ins_RM) particles had cubic shape and a size around 50 μm
alone, was employed. It uses a specific antibody immobilized on the (Fig. 1). Ins_SD powder comprises microparticles having a wrinkled
wells of the plate and an antibody labelled with the enzyme HRP shape. Shriveled microparticles derived from the collapse of protein
(horseradish peroxidase). The kit contains a diluted washing solution solution inflated droplets during drying. Ung et al. [21] previously
(Wash Buffer diluted 1:30 with distilled or deionized water), a substrate explored particle size and density of spray-dried insulin microparticles
solution (Substrate solution) and a stopping solution (Stop Solution). with the aim to maximize the total lung dose. The corrugated mor-
Results were expressed at each time point as mean ± standard phology, size and density of the present particles are consistent with the
error for each treatment. Glucose and insulin profiles were analyzed formulation properties of spray-dried insulin adopted in this study [21].
using two-way ANOVA followed by Bonferroni's post-test. Cmax, tmax Afrezza presented the typical microcrystalline plates structure of
and AUC0–120 were calculated using GraphPad Prism 8.3 and one-way Technosphere particles [15].
ANOVA followed by Tuckey's post-test was applied for statistical com- The powders Ins_SD and Afrezza were characterized in terms of
parison of groups. P < .01 were considered highly significant, aerodynamic performance (Table 2). Both products, Afrezza and Ins_SD,
p < .001, extremely significant. Taking Afrezza as reference, the showed an emission from the device > 90% and a remarkable respir-
Relative Bioavailability (RBA) was calculated as: ability. However, Ins_SD provided a superior performance presenting a
Afrezza dose MMAD of 0.9 μm, an FPF up to 91.5%, and an extra fine fraction of
Ins _SD AUC
RBA = × × 100 76%. The insulin aerodynamic distributions in the NGI are shown in
Afrezza AUC Ins _SDdose
Fig. 2. Reduced powder deposition in the induction port and increased
The RBA standard deviation of the ratio was calculated using the deposition from stages 5 to MOC were obtained with Ins_SD (see
Taylor expansion. Fig. 2a). The low amount of Ins_SD powder deposited into IP decreases
the risk of microparticles' deposition in the extrathoracic region. In
2.10. Bronchoalveolar lavage fluid (BALF) analysis and lung histology addition, the MOC captured more than 20% of the aerosolized particles,
which justifies the low value of MMAD. Their small aerodynamic size
After euthanasia bronchoalveolar lavage fluid (BALF) was retrieved leads to a deep lung deposition of the particles, without disregarding a
with three injections of 4 ml of HBSS containing 1 mM Hepes, 10 mM certain risk of particle exhalation. Since the process of drug release and
EDTA, centrifuged at 1800 rpm for 7 min at 4 °C and cells were sus- absorption are at their maximum level in the alveoli, the deep deposi-
pended in cell staining buffer (PBS containing 0.5% FCS and 0.1% so- tion would be effective for the systemic insulin uptake.
dium azide). Cellular suspension was incubated with IgG1-Fc (1 μg/106 Very interestingly, the aerodynamic properties of Ins_SD were not
cells) for 10 min at 4 °C, to block non-specific binding sites before affected by the product manufacturing process: MMAD and fine particle
staining with Phycoerythrin-Cyanine 5 (PE-Cy5) conjugated anti-Rat fraction measured before and after the semi-automatic filling proce-
CD45, 0.25 μg/106 cells (BD Biosciences™, San Jose, CA, US), dure, were similar. Therefore, the mechanical manipulation of powder
Fluorescein Isothiocyanate (FITC) anti-Rat CD3, 0.5 μg/106 cells, applied for metering the loaded dose did not affect the aerosolization
Phycoerythrin (PE) anti-Rat CD43, 0.5 μg/106 cells (BioLegend™, San behavior. In fact, the powder showed a favorable flowability during the

415
E. Quarta, et al. Journal of Controlled Release 323 (2020) 412–420

20–22%. This value is the half of the value measured for Ins_SD.
However, the distribution was not affected by the amount of powder
loaded in the DreamBoat inhaler. Finally, the NGI deposition profile of
Afrezza was positioned on the higher cut-off stages.

3.1. Stability study of Ins_SD

One of the major issues of insulin formulation is the chemical sta-

bility of the molecule that often requires the supply and storage of the
medical product using a cold chain. Afrezza has to be stored in re-
frigerated conditions and the powder cartridges has to be used within
3 days from the blister opening.
The USP42 monograph for injectable insulin preparations monitors
insulin stability by determining the formation of compound having
molecular masses greater than that of insulin (HMWP) and the forma-
tion of A21-desamido-insulin and other related proteins (ORP). The
other impurities with molecular masses greater than insulin have to
remain within the limit of 2.0%. The limit for the content of A21 is set
below 5.0% and that for related compounds below 6.0%. The trends of
the degradation products are reported in Fig. 3A and B. The percentage
of all Ins_SD decomposition products (A21, ORP and HMWP) was below
the USP limits (dotted line in the graphs) in both storage conditions for
the 6 months of investigation.
A stability test was performed also for the respirability of the
powder. Fig. 3C and D show the aerodynamic parameters of Ins_SD
stored at 4 °C and at 25 °C-60%RH. The results obtained allow to con-
clude that the respirability of Ins_SD powder remained unmodified
during the 6 months of investigation. The powder, stored both at room
temperature or in refrigerated conditions, maintained the high respir-
ability performance showed at time zero.
In summary, Ins_SD in combination with Quali-V®-I HPMC capsules,
specifically developed for use in dry powder inhalers and having a low
moisture content in the range 4.5–6.5%, provided a stability of the
product at room temperature up to six months. Therefore, while for
distribution it may be advisable to store the product in a cold chain,
once it reaches the patient's hands it could be kept at room temperature.

3.2. Pharmacokinetics and pharmacodynamic studies

With the aim to administer the dose of 10 IU/kg to the rats, Ins_SD
and Afrezza powders were blended with a carrier constituted by a
spray-dried mannitol powder, to obtain a sufficient mass to accurately
load the drug dose in the device. The blends showed a homogeneous
Fig. 1. Scanning electron microscopy images of insulin raw material (Ins_RM), insulin content; the drug content in the blends of spray-dried mannitol
pure insulin spray-dried powder (Ins_SD) and Afrezza.
with Ins_SD or Afrezza are reported in Table 3. Fig. 4 shows the aspect
of these insulin–mannitol blends to administer to rats by insufflation. In
filling process in the HPMC capsule for inhalation (Quali-V®-I). The both pictures the particles of insulin and the spherical particles of
filled capsule had a uniform insulin content of 1.78 ± 0.08 mg. mannitol can be clearly distinguished. The blend containing Afrezza
Afrezza was tested for comparison, emitting the powder from its (Afrezza/mann) (Fig. 4A) presents several typical Technosphere mi-
DreamBoat device. The powder mainly deposited on the stages 3, 4 and croplate particles, whereas Ins_SD shows the characteristics shriveled
5 of the impactor (Figs. 2B) and particles exhibited MMAD values be- and corrugated spray-dried particles. As seen (Table 3), the insulin
tween 3.1 and 3.2 μm. The 8 units and 4 units cartridges had a similar concentration in the blend was about 4% (w/w). In the SEM image of
deposition: the fine particle fraction lower than 5 μm was around Afrezza/mann blend the amount of mannitol particles seems not to
68–71% and the extra-fine particle fraction < 2 μm was around match with this concentration as Afrezza particles are much more

Table 2
Aerodynamic parameters of Ins_SD for the capsule filled manually or by semi-automatic filling process and of Afrezza (n = 3, mean value ± standard deviation).
Insulin loaded dose: Ins_SD 2 mg upon semi-automatic and 3 mg upon manual filling. Afrezza 8 units and 4 units cartridge contain 0.7 and 0.35 mg of insulin,
respectively.
Metered dosea (mg) Emitted Dose (mg) MMAD (μm) FPD (mg) FPF (%) Extra-FPD < 2 μm (mg) Extra-FPF < 2 μm (%)

Ins_SD semi-automatic 1.62 ± 0.08 1.50 ± 0.05 0.89 ± 0.09 1.37 ± 0.04 91.5 ± 5.9 1.14 ± 0.04 76.1 ± 5.3
Ins_SD manually 2.35 ± 0.06 2.26 ± 0.07 0.85 ± 0.09 2.07 ± 0.04 91.5 ± 1.2 1.73 ± 0.05 76.9 ± 0.2
Afrezza Cart. 8UI 0.64 ± 0.00 0.60 ± 0.00 3.19 ± 0.03 0.41 ± 0.01 68.3 ± 0.6 0.12 ± 0.01 20.8 ± 0.9
Afrezza Cart. 4UI 0.29 ± 0.01 0.28 ± 0.01 3.11 ± 0.02 0.20 ± 0.01 71.0 ± 0.5 0.07 ± 0.02 21.5 ± 0.1

a
Metered dose: insulin amount collected in the inhaler and deposited in the NGI impactor (induction port, stages 1 to 7 and MOC) quantified by HPLC.

416
E. Quarta, et al. Journal of Controlled Release 323 (2020) 412–420

Fig. 2. In vitro aerodynamic insulin distribution on different portions of the NGI of A) Ins_SD in manually and semi-automatically filled capsules and B) Afrezza
powder at the 8 units and 4 units strengths; Dev = device, IP = induction port, MOC = micro-orifice collector.

abundant than mannitol ones. This is correct though because Ins_SD small particle size and the absence of excipients in Ins_SD would lead to
and Afrezza powders have very different insulin content, i.e., 98% and a fast dissolution rate of peptide microparticles and, consequently,
18% (w/w), respectively. absorption rate. Afrezza is constituted by composite microparticles,
Both blends were successfully emitted (> 96%) by the DP-4 insuf- where insulin is embedded in a matrix of fumaryl-diketopiperazine,
flator (Penn Century), indicating that powders were not cohesive. possibly affecting the dissolution time, macrophage phagocytosis or
Finally, the dv,50 of the emitted powder plumes were measured by laser ciliary transport.
diffraction. The volume diameter value of Ins_SD blend (Ins_SD/mann) Taking in consideration IT and SC administration, the difference in
in Table 3, compared with the previous reported volume diameters of exposure is due to the intrinsic differences between the two adminis-
Ins_SD (2.7 μm) and spray-dried mannitol (1.9 μm), indicates that the tration types. It is known that during the pulmonary administration a
air flow produced by the insufflator could properly deaggregate the fraction of the dose can be lost in the device, whereas the SC admin-
formulation (Table 3). Summarizing, the collected data demonstrate istration the unit of insulin are quantitatively injected in the tissue [3].
that the insufflator was suitable to aerosolize the blends for the animal For this reason, a clinical study was conducted with Afrezza to find the
IT administration. dose correlation between the two routes of administration It was de-
Ins_SD/mann given intratracheally, was promptly absorbed, termined that 8 units of ultra rapid insulin lispro injected sub-
reaching the peak plasma concentration (Cmax 4.9 ± 1.5 mU/ml) cutaneously were PK/PD equivalent to Afrezza 8 units strength (con-
15 min post-dosing (Fig. 5). The concentration at 120 min after ad- taining 20 IU of insulin) [24] (web ref 2).
ministration was negligible (0.03 mU/ml). Afrezza/mann showed
slower absorption rate (tmax 30 min), lower Cmax (1.8 ± 0.37 mU/ml) 3.3. Pharmacodynamics studies
and a similar concentration at 120 min.
Subcutaneous insulin showed slow absorption rate (tmax 60 min), The rat model is suitable for the PD study due to the efficacy of
reaching a non-significantly different peak of insulin plasma con- inhalation administration. The pharmacodynamic profile of Ins_SD in
centration compared to Ins_SD/mann intratracheally administered rats consisted in measuring the glucose plasma profile after in-
(p = .7). As expected, insulin AUC0–120 after subcutaneous adminis- traperitoneal injection of a bolus of 1 g/kg glucose followed by IT or SC
tration was the highest. In fact, AUC0–120 of subcutaneous insulin insulin administration. As shown in Fig. 7, all experimental groups had
(352 ± 38 mU/ml ∗ min) was significantly higher than AUC0–120 of similar initial basal glucose plasma levels (150 mg/dl). Following glu-
intratracheal Ins_SD/mann administration (144 ± 29 mU/ml ∗ min) cose administration, glycaemia rapidly reached the concentration of
and of Afrezza/mann (78 ± 14 mU/ml ∗ min) (Fig. 6). Albeit the dif- 430 ± 21 mg/dl in 30 min, as it can be seen in control rats inhaling
ference in AUC0–120 of Afrezza/mann and Ins_SD/mann was not sta- mannitol microparticles (Mann_SD). Glucose values achieved for this
tistically significant, the latter presented an RBA equal to 185% ± 156. experimental group were significantly higher than basal values even
The longer exposure to insulin can be assigned to the slow insulin after 180 min and the curve profile was consistent with previous find-
release and the total dose injected available for absorption. The data are ings [25].
in agreement with previous studies showing that dry powder insulin for All insulin formulations (Ins_SD/mann, Afrezza/mann inhalation
inhalation had a faster absorption, but lower bioavailability relative to powder and subcutaneous insulin) prevented the raise of glycaemia
SC injection [6,22,23]. after glucose administration. Any insulin treatment kept plasma glucose
Profiles in Fig. 5 showed that Ins_SD had a more rapid absorption of concentrations significantly lower compared to the values obtained
the peptide compared to Afrezza/mann. In fact, a significantly higher after mannitol powder inhalation, over the whole observation time.
Cmax (p < .01) was measured. Considering that the administered dose Rats treated with insulin SC, Afrezza/mann and Ins_SD/mann
was the same (10 IU/kg), the higher absorption rate (tmax and Cmax) of showed a similar initial glycemic small peak after 15 min and glycemic
Ins_SD/mann in comparison to Afrezza/mann could be explained firstly curves for 60 min after glucose administration. Plasma glucose of IT rats
by the higher fine particle dose leading to a deeper deposition of Ins_SD reverted to basal conditions after 2 h, whereas a prolonged decrease
particles, where the drug absorption is maximized. Indeed, Ins_SD was observed for SC insulin. This profile agreed with insulin plasmatic
presented a more favorable respirability due to an FPF of 91% deriving concentration due to a rapid absorption (tmax 15 min) for IT treatment
from a very small aerodynamic size (MMAD 0.9 μm). Secondly, the and the longer insulin exposure for subcutaneous administration.

417
E. Quarta, et al. Journal of Controlled Release 323 (2020) 412–420

Fig. 3. Degradation products of insulin spray-dried powder Ins_SD stored in extra-dry Quali-V®-I capsules and Alu-Alu blister at room temperature (25 °C, 60% RH)
and at 4 °C for six months: (A) A21 desamido insulin + other related proteins (ORP); (B) high molecular weight protein (HMWP) degradation products (USP42 limits
are represented by the dotted lines). Aerodynamic parameters of Ins_SD upon aerosolization with RS01_HR inhaler: Emitted Dose (mg) (C); Fine Particle Dose < 5 μm
(mg) (D); (mean ± standard deviation, n = 3).

3.4. Histological analysis and flow cytometry arrows in the pictures of Fig. 8). In summary, histologically significant
differences between the control (Mann_SD) and the 3 experimental
The histological studies and BALF were used to evaluate the possible groups did not emerge with respect to pulmonary parenchyma samples.
inflammatory effects of the drug formulation for inhalation. Fig. 8 il- Beside the histological analysis, the percentage of granulocyte and
lustrates the histological analysis of lung tissue after administration of monocyte in the BALF was determined to evaluate the lung in-
the different IT or SC insulin formulations. In all samples, moderate flammation upon the insulin administration (Fig. 9).
pulmonary vessel hyperaemia phenomena were observed in pulmonary Here, in the animals that received intratracheally just an air puff,
parenchyma specimens; rarely, moderate mild alveolar edema was the percentage of inflammatory cells (granulocytes and monocytes) was
detected as well. Infiltration of polymorphonuclear cells into the al- around 15%. This value was in line with the other groups receiving
veolar wall and perivascular tissue was also observed (indicated by intratracheal administrations. In contrast, animals treated with SC

Table 3
Insulin content and dv,50 of insulin/mannitol powder blends upon emission using DP-4 insufflator dry powder inhaler (n = 4, mean ± standard deviation). Expected
insulin content in the blend: 4% (w/w).
Insulin content (%, w/w) Emitted dose by DP-4 insufflator Penn Century (%) dv,50 (μm)

Ins_SD/mann 4.47 ± 0.09 96.3 ± 0.5 2.13 ± 0.10

Afrezza/mann 4.06 ± 0.06 97.2 ± 0.3 3.57 ± 0.12

418
E. Quarta, et al. Journal of Controlled Release 323 (2020) 412–420

Fig. 4. Scanning electron microscopy pictures of Afrezza/mann (A) and Ins_SD/mann (B) powder blends.

Fig. 7. Pharmacodynamic profile. Plasma glucose concentrations after 1 g/kg

Fig. 5. Insulin plasma concentrations in rats receiving different insulin for-
mulations. Data are the mean ± SEM of 9 rats. Two-way ANOVA followed by glucose intraperitoneal injection followed by insulin administration. Data are
Bonferroni's post-test §§§ = p < .001 (Ins_SD/mann vs Afrezza/mann) the mean ± SEM of 9 rats. Two-way ANOVA followed by Bonferroni's post-
test. ** = p < .01; *** = p < . 001 (Ins_SD/mann vs Mann_SD).
### = p < .001 (Ins_SD/mann vs Ins SC).
## = p < .01; ### = p < .001 (Ins_SD/mann vs Ins SC).

Fig. 6. Mean AUC0–120 of Ins_SD/mann, Afrezza/mann and Insulin SC (10 UI/

kg). Data are the mean ± SEM of 9 rats. One-way ANOVA followed by Tukey's
post-test *** = p < .001.

showed only an increase of about 3% of granulocyte and monocytes.

The difference between SC and intratracheal groups was significant in
all the cases (p < .01). Considering these results, it was concluded that Fig. 8. Lung tissue 3 h post administration of the different formulations (mag-
the lung inflammation was to be attributed to the mechanical insertion nification: 20×, hematoxylin-eosin). A: Mann_SD; B: Insulin SC; C: Ins_SD/
of the device needle and not to the administered formulations. mann; D: Afrezza/mann.

a very high respirability. Furthermore, insulin spray-dried powder

4. Conclusions
stability data have shown that the extra-dry HPMC capsules packaged
in an Alu-Alu blister can provide a product stability at room tempera-
Pulmonary administration of Ins_SD and Afrezza powders blended
ture up to six months, preserving good aerodynamic performance.
with a mannitol microparticulate carrier promptly reduced the gly-
These findings open the possibility to execute the daily therapy without
cemia level, exhibiting less pronounced hypoglycaemia than after in-
worring about the refrigeration of the product.
sulin subcutaneous injection.
Ins_SD/mann intratracheally was promptly absorbed, reaching a
Pure insulin pulmonary powder, produced by spray drying from an
maximal concentration 15 min post-dosing, whereas Afrezza/mann
acidic aqueous solution of peptide and filled in hard capsules, presented

419
E. Quarta, et al. Journal of Controlled Release 323 (2020) 412–420

insulin, Clin. Ther. 36 (2014) 1275–1289, https://doi.org/10.1016/j.clinthera.

2014.06.025.
[6] L. Heinemann, R. Baughman, A. Boss, M. Hompesch, Pharmacokinetic and
Pharmacodynamic properties of a novel inhaled insulin, J. Diabetes Sci. Technol. 11
(2017) 148–156, https://doi.org/10.1177/1932296816658055.
[7] B.W. Bode, J.B. McGill, D.L. Lorber, J.L. Gross, P.-C. Chang, D.B. Bregman, et al.,
Inhaled technosphere insulin compared with injected prandial insulin in type 1
diabetes: a randomized 24-week trial, Diabetes Care 38 (2015) 2266–2273, https://
doi.org/10.2337/dc15-0075.
[8] F. Andrade, P. Fonte, A. Costa, C.C. Reis, R. Nunes, A. Almeida, et al.,
Pharmacological and toxicological assessment of innovative self-assembled poly-
meric micelles as powders for insulin pulmonary delivery, Nanomedicine (London)
11 (2016) 2305–2317, https://doi.org/10.2217/nnm-2016-0045.
[9] K. Rave, T. Heise, L. Heinemann, A.H. Boss, Inhaled technosphere insulin in com-
parison to subcutaneous regular human insulin: time action profile and variability
in subjects with type 2 diabetes, J. Diabetes Sci. Technol. 2 (2008) 205–212,
https://doi.org/10.1177/193229680800200206.
[10] J. Pettus, T. Santos Cavaiola, S.V. Edelman, Recommendations for initiating use of
Afrezza inhaled insulin in individuals with type 1 diabetes, Diabetes Technol. Ther.
20 (2018) 448–451, https://doi.org/10.1089/dia.2017.0463.
[11] L.N. Patel, J. Wang, K.-J. Kim, Z. Borok, E.D. Crandall, W.-C. Shen, Conjugation
Fig. 9. Percentage of granulocyte and monocyte (CD43+ cells; Gated on with cationic cell-penetrating peptide increases pulmonary absorption of insulin,
CD45+ CD3-) in the BALF after intratracheal vs subcutaneous insulin (Ins SC) Mol. Pharm. 6 (2009) 492–503, https://doi.org/10.1021/mp800174g.
administration. Data are the mean ± SEM of 9 rats. One-way ANOVA followed [12] J. He, G. Zhang, Q. Zhang, J. Chen, Y. Zhang, X. An, et al., Evaluation of inhaled
by Dunnett's post-test ** = p < .01. recombinant human insulin dry powders: pharmacokinetics, pharmacodynamics
and 14-day inhalation, J. Pharm. Pharmacol. 71 (2019) 176–184, https://doi.org/
10.1111/jphp.13026.
showed a slower absorption rate. Concerning the glycemia control, [13] L. Garcia-Contreras, T. Morçöl, S.J.D. Bell, A.J. Hickey, Evaluation of novel particles
as pulmonary delivery systems for insulin in rats, AAPS PharmSci. 5 (2003) 10–20,
Ins_SD/mann provided a rapid reduction of plasma glucose levels si- https://doi.org/10.1208/ps050209.
milar to Afrezza/mann, which lasted shorter compared to SC injection. [14] J.G. Weers, D.P. Miller, T.E. Tarara, Spray-dried PulmoSphere™ formulations for
The Ins_SD product is suitable for a post-prandial glucose control pro- inhalation comprising crystalline drug particles, AAPS PharmSciTech 20 (2019)
103–115, https://doi.org/10.1208/s12249-018-1280-0.
viding the patient with a convenient and compliant product, in parti- [15] E. Potocka, J.P. Cassidy, P. Haworth, D. Heuman, S. van Marle, R.A. Baughman,
cular in the event of using a disposable device.. Pharmacokinetic characterization of the novel pulmonary delivery excipient fu-
The insulin Ins_SD product has to be stored in refrigerated container maryl diketopiperazine, J. Diabetes Sci. Technol. 4 (2010) 1164–1173, https://doi.
org/10.1177/193229681000400515.
during distribution to ensure shelf-life stability. The six months of the
[16] P.C. Richardson, A.H. Boss, Technosphere insulin technology, Diabetes Technol.
room temperature stability found makes less dependent on storage in Ther. 9 (Suppl. 1) (2007) S65–S72, https://doi.org/10.1089/dia.2007.0212.
refrigerated conditions the transport and use of the product once in the [17] L. Heinemann, C.G. Parkin, Rethinking the viability and utility of inhaled insulin in
clinical practice, J. Diabetes Res. 2018 (2018) 4568903–4568909, https://doi.org/
patient's hands.
10.1155/2018/4568903.
This work demonstrated that an active insulin pulmonary powder, [18] Cagnani, S., Colombo, P., Ventura, P., 2004. Insulin highly respirable micro-
having high respirability was successfully prepared by spray drying particles. Assignee: University of Parma; US Patent Number: 7625865 B2.
without excipients. Finally, the regulatory agencies would be less [19] A.G. Balducci, S. Cagnani, F. Sonvico, A. Rossi, P. Barata, G. Colombo, et al., Pure
insulin highly respirable powders for inhalation, Eur. J. Pharm. Sci. 51 (2013)
concerned about the safety and tolerability with this excipient-free 110–117, https://doi.org/10.1016/j.ejps.2013.08.009.
pulmonary formulation. [20] S.J. Nolan, J. Thornton, C.S. Murray, T. Dwyer, Inhaled mannitol (Bronchitol) for
cystic fibrosis, Paediatr. Respir. Rev. 18 (2016) 52–54, https://doi.org/10.1016/j.
prrv.2015.12.001.
Acknowledgment [21] K.T. Ung, N. Rao, J.G. Weers, D. Huang, H.-K. Chan, Design of spray dried insulin
microparticles to bypass deposition in the extrathoracic region and maximize total
The authors would like to thank Plastiape Spa (Osnago, LC, Italy) for lung dose, Int. J. Pharm. 511 (2016) 1070–1079, https://doi.org/10.1016/j.
ijpharm.2016.07.073.
kindly supplying the RS01 inhaler and Qualicaps for partially funding [22] M. Amidi, K.M. Krudys, C.J. Snel, D.J.A. Crommelin, O.E. Della Pasqua,
the research project. Authors would like to acknowledge Cataldo W.E. Hennink, et al., Efficacy of pulmonary insulin delivery in diabetic rats: use of a
Martucci, Franco Bassani and Fabio Stellari, personnel of Chiesi model-based approach in the evaluation of insulin powder formulations, J. Control.
Release 127 (2008) 257–266, https://doi.org/10.1016/j.jconrel.2008.01.019.
Farmaceutici Spa, for the training and support in the execution of in vivo
[23] M. Gopalakrishnan, S. Suarez, A.J. Hickey, J.V.S. Gobburu, Population pharmaco-
study. kinetic-pharmacodynamic modeling of subcutaneous and pulmonary insulin in rats,
J. Pharmacokinet. Pharmacodyn. 32 (2005) 485–500, https://doi.org/10.1007/
s10928-005-0008-4.
References
[24] A.H. Boss, R. Petrucci, D. Lorber, Coverage of prandial insulin requirements by
means of an ultra-rapid-acting inhaled insulin, J. Diabetes Sci. Technol. 6 (2012)
[1] T. Goldberg, E. Wong, Afrezza (insulin human) inhalation powder: a new inhaled 773–779, https://doi.org/10.1177/193229681200600406.
insulin for the management of Type-1 or Type-2 diabetes mellitus, P T. 40 (2015) [25] K. Yassin, V.T.T. Huyen, K.N. Hoa, C.G. Ostenson, Herbal extract of gynostemma
735–741. pentaphyllum decreases hepatic glucose output in type 2 diabetic goto-kakizaki
[2] A.H. De Boer, P. Hagedoorn, M. Hoppentocht, F. Buttini, F. Grasmeijer, rats, Int. J. Biomed. Sci. 7 (2011) 131–136.
H.W. Frijlink, Dry powder inhalation: past, present and future, Expert Opin. Drug
Deliv. 14 (2017) 499–512, https://doi.org/10.1080/17425247.2016.1224846.
[3] J. Patton, J. Bukar, S. Nagarajan, Inhaled insulin, Adv. Drug Deliv. Rev. 35 (1999)
Web references
235–247, https://doi.org/10.1016/s0169-409x(98)00074-x.
[4] J. Oleck, S. Kassam, J.D. Goldman, Commentary: why was inhaled insulin a failure [1] https://aerami.com/clinical-studies/#.
in the market? Diabetes Spectr. 29 (2016) 180–184, https://doi.org/10.2337/ [2] https://www.accessdata.fda.gov/drugsatfda_docs/label/2014/022472lbl.pdf.
diaspect.29.3.180.
[5] T.S. Cavaiola, S. Edelman, Inhaled insulin: a breath of fresh air? A review of inhaled

420