Shock Waves DOI 10.1007/s00193-013-0464-5

ORIGINAL ARTICLE

A new biolistic intradermal injector M. Brouillette · M. Doré · C. Hébert · M.-F. Spooner · S. Marchand · J. Côté · F. Gobeil · M. Rivest · M. Lafrance · B. G. Talbot · J.-M. Moutquin

Received: 22 June 2012 / Revised: 21 February 2013 / Accepted: 23 February 2013 © Springer-Verlag Berlin Heidelberg 2013

Abstract We present a novel intradermal needle-free drug delivery device which exploits the unsteady high-speed flow produced by a miniature shock tube to entrain drug or vaccine particles onto a skin target. A first clinical study of pain and physiological response of human subjects study is presented, comparing the new injector to intramuscular needle injection. This clinical study, performed according to established pain assessment protocols, demonstrated that every single subject felt noticeably less pain with the needle-free injector than with the needle injection. Regarding local tolerance and skin reaction, bleeding was observed on all volunteers after needle injection, but on none of the subjects following powder injection. An assessment of the pharmacodynamics, via blood pressure, of pure captopril powder using the new device on spontaneously hypertensive rats was also performed. It was found that every animal tested with the needle-free injector exhibited the expected pharmacodynamic response following captopril injection. Finally, the new injector was used to study the delivery of an inactivated influenza vaccine in mice. The needle-free device induced serum antibody response Communicated by S. H. R. Hosseini. M. Brouillette (B) · M. Doré · C. Hébert Department of Mechanical Engineering, Université de Sherbrooke, SherbrookeJ1K 2R1, Canada e-mail: [email protected] M.-F. Spooner · S. Marchand · J. M. Moutquin Clinical Research Center, Université de Sherbrooke, Sherbrooke, Canada J. Côté · F. Gobeil Department of Pharmacology, Université de Sherbrooke, Sherbrooke, Canada M. Rivest · M. Lafrance · B. G. Talbot Department of Biology, Université de Sherbrooke, Sherbrooke, Canada

to the influenza vaccine that was comparable to that of subcutaneous needle injection, but without requiring the use of an adjuvant. Although no effort was made to optimize the formulation or the injection parameters in the present study, the novel injector demonstrates great promise for the rapid, safe and painless intradermal delivery of systemic drugs and vaccines. Keywords Drug delivery · Biolistics · Miniature shock tube

1 Introduction Intradermal powder injection is an emerging technology for the needle-free delivery of a potentially wide array of drugs and vaccines. Although needle injection of liquids is widespread principally because of its low cost, this delivery method is painful, generates dangerous medical waste and can cause contamination. Various technologies have been developed to address these shortcomings, amongst them creams, patches, inhalers and liquid jet injectors, each with their own severe limitations. Intradermal powder, i.e., biolistic, delivery has emerged from gene gun technology, in which DNA-coated metal particles (of the order of 1 μm diameter) are propelled at high impact speeds onto plant cells for genetic modification [1,2]. In intradermal powder injection, the particles containing the drug or vaccine are accelerated to sufficiently high speed to penetrate into targeted skin layers to achieve the desired pharmacological effect. Prototype devices, using pressurized gas [3], laser [4] or explosives [5], to accelerate the particles, have been shown to be applicable to the delivery of proteins [6], conventional vaccines [7] and DNA vaccines [8].

123

M. Brouillette et al.

We are developing a family of low-cost intradermal injectors applicable to the delivery of powder particles [9] as well as liquid droplets [10], for both systemic drug and vaccine applications. This article first presents the features of the powder delivery device from a gas dynamics point of view, and describes its technical capabilities. The main objective of the present study is to evaluate the applicability and performance of a laboratory version of our needle-free injector, with regard to pain/local tolerance in humans and both systemic drug and vaccine delivery in animals. In particular, we present the results of a pain, local tolerance and skin reaction study in human subjects, comparing this device with conventional intramuscular needle injection. We also demonstrate the capability of the powder injector to produce the desired pharmacological outcome with a systemic drug in a spontaneously hypertensive rat model. Finally, we demonstrate the capability of the powder injector to inject a vaccine and produce the expected immunological response.

2 Intradermal powder delivery device Our powder delivery device exploits the unsteady flow generated by the triggering of a miniature shock tube to entrain micron-scale particles at high speed onto a skin target. The device further exploits an area contraction between the driver and driven sections of the shock tube to produce higher gas, and thus particle, velocities than a constant area shock tube. This method of operation differs greatly from the device of Bellhouse et al. [3] which attempts to entrain the particles into a quasi-steady supersonic gas stream produced by a converging–diverging nozzle; however in this device, performance is impaired by the fact that nozzle start-up time is comparable to the particle ejection time [11]. Our device also differs from that of Kendall et al. [12], which uses a contoured shock tube with a divergent exit to achieve pressure-matched operation. Our device is shown schematically in Fig. 1. A reservoir is initially filled with driver gas at elevated pressure: here nitrogen at 32 bar. This reservoir is connected to the shock tube driver section with a quick-opening valve. The 19-mm-long driver is connected to the test section through a 9-to-1 area contraction and the driven section is straight with an inner diameter of 6 mm and a length of 51 mm; the driver and driven sections are initially filled with atmospheric air and separated by a double membrane arrangement with the powder dose sandwiched in between. The flow exits the driven section into a larger silencing chamber and interacts with the skin target through an opening. A cross section of a reusable/autoclavable laboratory version of the device is shown in Fig. 2 and a photograph is shown in Fig. 3.

123

Fig. 1 Schematic of a needle-free powder delivery device

Fig. 2 Cross section of a reusable/autoclavable laboratory version of the needle-free powder delivery device. The device is fabricated mostly out of 316 stainless steel. The check valve is used to fill the reservoir before each test

Fig. 3 Photograph of a reusable/autoclavable laboratory version of the needle-free powder delivery device. The device of Fig. 2 is here covered with an an ergonomic plastic shell

The device is triggered by opening the reservoir into the driver, which raises the driver pressure to the burst pressure of the membranes (13 μm polyester film) at 10.8 bar. At membrane rupture, an unsteady flow is established in the device, as shown in the wave diagram of Fig. 4. As in a conventional shock tube, a shock is produced in the driven section, while unsteady expansions (primary expansion) are produced in the driver. Because of the area contraction between the driver and driven sections, the subsonic driver gas is further accelerated almost isentropically through the convergent. If the area or pressure ratios are sufficiently large, the convergent becomes choked and further flow acceleration is achieved via left-facing unsteady expansion waves (secondary expansion) that are swept downstream into the now supersonic flow region. The particles, initially at rest at the outlet of the contraction, are accelerated by the high-velocity, high-density driver gas until their interaction with the reflected expansion waves from the driver end wall. At the exit of the driven

A new biolistic intradermal injector

Fig. 4 Quasi 1-D wave diagram of unsteady flow phenomena in a needle-free powder delivery device

section, the shock wave reflects as expansion waves which further accelerate the particles. Since the flow at the nozzle exit is initially underexpanded, a steady expansion (not shown) further takes place just outside the nozzle, which accelerates the particles to their maximum velocity. Finally, a normal shock wave stands off a certain distance from the skin target (not shown), reducing the impact velocity. The particle velocities can be obtained in the free field, i.e., without the silencer and target, using particle image velocimetry (PIV). The experimental setup consisted of a New Wave Research Solo PIV twin-head Nd:Yag laser system emitting at 532 nm paired with a Lavision PIV system. Both images were recorded at a 1 μs interval through a Navitar 12× zoom lens attached to a Lavision Imager Pro X 4MP camera with a 2,048 × 2,048 pixels resolution. The resulting shots were analyzed using the DaVis 7.2 software from Lavision: to determine the vector flow field, a multi-pass scheme was used with a first pass square interrogation window of 64 × 64 pixels. This vector field was then used as an initial velocity estimate for a second pass using a 32 × 32 window. This process was repeated again twice with a 16 × 16 window, once to get another velocity estimate, and the final pass to get the correct flow field. All the cross-correlations performed to determine the vector fields (estimates and final) were done with a 50 % window overlap. Figure 5 shows the horizontal flow velocity field obtained at the above test conditions with a 2 mg particle dose of pure trehalose sieved to 45–53 μm. The flow was captured 250 μs after the shock crossed a pressure transducer located 15 mm downstream of the membrane. Particle velocities between 300 and 500 m/s are achieved with the device, which are sufficient to breach the stratum corneum layer of most skin targets [13]. Although of small scale, this miniature shock tube is sufficiently large to avoid the deleterious effects of scale outlined by Brouillette [14] for microscale shock tubes.

Fig. 5 Horizontal flow velocity field with a 2 mg particle dose of pure trehalose sieved to 45–53 μm. The flow was captured 250 μs after the shock crossed a pressure transducer located 15 mm downstream of the membrane. Nitrogen driver burst pressure was 10.8 bar

3 Pain and local tolerance study Here we present the results of a pain, local tolerance and skin reaction study in human subjects, comparing this device with conventional intramuscular needle injection. This study is performed according to recognized pain assessment protocols on human subjects. In particular, we compare needlefree injections of sodium chloride, in pure powder form, with intramuscular injections of a similar sodium chloride solution.

3.1 Methods 3.1.1 Reagents The following reagents were used: sodium chloride ACS grade (EMD Chemicals) and sterile isotonic saline solution 0.9 % (Abbott Laboratory).

3.1.2 Subjects Subjects were men and women between 18 and 40 years old. Any volunteers suffering from acute, chronic or diffuse pain pathology were excluded from the study. Subjects with cardiac history, elevated blood pressure, pulmonary problems or taking vasoactive medication were also excluded. It was not recommended to smoke or drink coffee 1 h before testing or take any medications that could alter pain perception. Menstrual cycle phases were not taken into consideration for this test.

123

M. Brouillette et al.

3.1.3 Protocol The study was performed with ten subjects (n = 10) by injecting pure sodium chloride powder with the powder injector and comparing it with an intramuscular needle injection of a saline solution. The experiment proceeded as follows. First, the difference between pain intensity and discomfort was described to subjects according to standard scales (discussed below). Electrocardiogram and skin conductance transducers were then attached to the volunteers. Before the injections, baseline physiological parameters were recorded. A questionnaire to evaluate their anxiety (IASTA form Y-2) was also completed by all volunteers. Following this, the powder and needle injectors were presented and an empty powder injector was applied to the forearm of the subject without triggering, to demonstrate how it worked. Another anxiety assessment was also performed at this point (IASTA form Y-1). The study was performed as a randomized crossover experiment. For each subject, the needle syringe (Terumo, 3 cc/ml, 21 gauge × 1−1/2”) contained 1 ml of sterile saline solution and the powder injector contained 2 mg of pure sterilized sodium chloride powder. An injection was first performed on the left or right deltoid (random) with either powder or needle injector (random), then another injection was performed on the other deltoid with the other injector; the second injection was performed about 10 min later on the other arm. Needle injections were intramuscular and powder injections were intradermal. After each injection, the volunteer rated a standard visual analog scale for pain intensity and for the uncomfortable aspect of the injection. These scales are graduated from 0 to 100 (0 = no pain at all/no discomfort, 100 = worst pain imaginable/worst discomfort). Physiological parameters (heart rate (HR), skin conductance (GSR)) were recorded throughout the experiment. Blood pressure and cutaneous reactions were noted at the end of each injection and cutaneous reactions were also evaluated 1 and 5 min following each injection. At the end of the experiment, each subject was asked which injection they preferred. The protocol was approved by the Human Research Ethics Committee at the Université de Sherbrooke. 3.1.4 HR measurements Measurements of HR variations (HRV) gives reliable information on the involvement of sympathetic and parasympathetic systems and overall autonomic influences affecting HR [15]. Also, real-time HR measurements are used to index autonomic responses. In fact, it has been established that the sympathetic nervous system is the main contributor to the rise of HR [16]. Electrocardiogram (ECG) records were made at a sampling frequency of 1,000 Hz with an ADInstruments Pow-

123

erlab monitor. This apparatus used low-pass and high-pass filters set at 1 kHz and 0.3 Hz, respectively, combined with a 60 Hz notch filter. The normal-to-normal (NN) interval is obtained by measuring the time between successive QSR wave complexes sampled by the ECG monitor. HRV software from ADInstruments was used to analyze the interbeat intervals. At 2 min baseline and at 15 s after injections, the mean HR was calculated [17]. High-frequency power (HF 0.15 to 0.40 Hz) and low-frequency power (LF 0.04 to 0.15 Hz) were also calculated; LF represents sympathetic and parasympathetic modulation and HF represents primarily vagal (parasympathetic) regulation of HR. Parasympathetic variations in HR were estimated from the number of NN interval differences >50 ms (NN50). LF/HF ratio is considered as a measurement of sympathovagal balance [15]. 3.1.5 GSR measurements Electrodermal measurements have been used to capture peripheral sympathetic response via sweat gland activity. GSR values were measured with another Powerlab monitor; no filter was used. Electrodes were affixed to the clean skin of the right hand of the subject (thenar and hypothenar eminence). Subjects had to keep their hand still to minimize experimental artifacts. Baseline conductivity measurements were recorded before each test. 3.1.6 Sodium chloride powder injections NaCl powder was ground with a pestle and mortar and then sieved to retain particle diameters within a 20 to 45 μm range using stainless steel sieves (Tyler). This material was then used to prepare 2 mg doses for the powder injector. Doses were sterilized in a glass container in an oven at 180◦ C for 2 h [18]. 3.2 Results 3.2.1 Pain evaluation Results show that the anticipated pain, the pain sensation and patient discomfort are significantly lower with the powder injector than with the needle (Fig. 6). Also, after the injections, the pain experienced lasted much longer after needle injection than with the powder injector, where it was essentially pain-free immediately thereafter. 3.2.2 Heart rate and blood pressure HR results did not exhibit any significant difference between needle and powder injection: Fig. 7 shows that the needle and powder injections brought similar and clinically insignificant variations in heart rate. It was also observed that the two injec-

A new biolistic intradermal injector 50

30

Powder injector Needle syringe

500

Delta GSR (%)

40

Pain scale (0-100)

600

Needle syringe (n=10) Powder injector (n=10)

20

400 300 200 100 0 Base

10

0-15

15-30

30-45

45-60

60-75

75-90 90-105 105-120

Time (sec) 0 Anticipated pain

Pain felt

Patient discomfort

Fig. 6 Pain evaluation after intramuscular needle injection (saline solution) and needle-free powder injection (pure sodium chloride) (n = 10)

4 2 0 -2

NN50

37,5

25

12,5

-4 -6 -8 0-15

0 Baseline 15-30

30-45

45-60

60-75

75-90

90-105 105-120

Time (sec)

Fig. 7 Variation of heart rate after intramuscular needle injection (saline solution) and needle-free powder injection (pure sodium chloride) (n = 10)

Powder injector

Needle syringe

Fig. 9 Parasympathetic parameters after injections. Comparison between intramuscular needle injection (saline solution) and needlefree powder injection (pure sodium chloride) (n = 10) 80

LF

5

LF/HF ratio

4

3.2.3 Galvanic skin response

60

Mean LF

tion methods did not significantly modify the blood pressure as compared to the baseline value.

3 40 2 20

Like the HR results, GSR did not vary between injection type. Figure 8 shows similar GSR curves with a fast peak of skin sweating right after the injections; the return to baseline takes about 2 min after the injections. These data demonstrate that the peripheral sympathetic system and central nervous system’s activation respond in the same way to both injection methods. 3.2.4 Parasympathetic and sympathetic system Analysis of electrocardiograms was performed for each subject. Since the HF and NN50 parameters are indicators of parasympathetic modulation, it can be noted that the parasympathetic system activity is slightly decreased with

1

Mean of LF/HF ratio

Heart rate variation (%)

6

HF

50

Indicator's mean

Powder injector Needle syringe

8

Fig. 8 Galvanic skin response for intramuscular needle injection (saline solution) and needle-free powder injection (pure sodium chloride) (n = 10)

0

0 Baseline

Powder injector

Needle syringe

Fig. 10 Sympathetic parameters after injections. Comparison between intramuscular needle injection (saline solution) and needle-free powder injection (pure sodium chloride) (n = 10)

both methods of injections as compared to the baseline data (Fig. 9). Figure 10 shows variation of sympathetic parameters, namely the LF and LF/HF ratios. It is observed that sympathetic system activity increased with both methods of injection.

123

M. Brouillette et al. Table 1 Skin reaction after needle injection

Subject

1 min after injection

5 min after injection

BC14

Light bleeding

–

MP15

Light bleeding

–

MAD16

Light bleeding, red spot on 3 cm

Small bleeding

HG17

Very light bleeding

–

GR18

Light bleeding, swelling on 3 mm

Still painful

GL19

Light bleeding, swelling on 3 mm

Still painful, swelling on 3 mm

MET20

Light bleeding

Still painful after 7 min

NH21

Light bleeding, red spot on 3 mm

–

NF22

–

Light bleeding, swelling on 3 mm, red spot on 3 cm

NC23

Light bleeding

–

SR24

Light bleeding

–

Table 2 Skin reaction after needle-free injection

1 min after injection

5 min after injection

BC14

–

–

MP15

–

–

MAD16

–

–

HG17

Light red spot on 3 cm

–

GR18

Light red spot

–

GL19

Very light red spot on 3 cm

–

MET20

Very light red spot on 3 cm

–

NH21

Very light red spot on 3 cm

Little “tickle” sensation after few minutes

NF22

Red spot on 3 cm, light red spot on 8 cm

–

NC23

Light red spot on 3 cm

–

SR24

Very light red spot

–

3.2.5 Injection site reaction

3.3 Discussion

Powder and needle injections produced different skin reactions from one volunteer to the next. Bleeding was observed for every subject after needle injection, and some swelling was also seen on some volunteers. Table 1 lists the skin reaction observed for each subject following the needle injection. For powder injection of NaCl, no bleeding was observed for any subject. Most subjects exhibited a small (<3cm) pink spot one minute after the injection, due to the increased local salt concentration brought upon by the powder injection. These localized skin reactions disappeared shortly after the injection, as local salt concentration stabilized to preinjection levels. Table 2 lists the skin reaction observed for each subject following the powder injection. However, this protocol did not involve the microcharacterization of potential localized damage to cells due to particle penetration and/or exposure to high-pressure gas.

Needle injection brought about three times more pain to the subjects than the powder injector. Similar differences were observed for the uncomfortable aspect of the needle syringe versus the powder injector. Cutaneous reaction from needle injection caused bleeding for all subjects. However, no bleeding was seen after powder injection. Only temporary small red spots, due to the salt injection, were observed on subjects following powder injection. Furthermore, the powder injector did not make a significant difference on the measured physiological parameters as compared to needle injection. Specifically, HR, GSR and blood pressure were very similar after both injections. Also, both injection methods were found to simultaneously increase sympathetic system response and decrease parasympathetic system response. In the end, the totality (100 %) of tested subjects preferred the needle-free powder delivery to the intramuscular needle injection.

123

A new biolistic intradermal injector

4 Systemic drug delivery Here, we demonstrate the capability of the powder injector to produce a desired pharmacological outcome with a systemic drug in a spontaneously hypertensive rat (SHR) model. We use captopril, an angiotensin-I converting enzyme inhibitor which is used in the treatment of hypertension and congestive heart failure. It is readily available in stable powder form and has the potential to be injected in pure doses, and transdermal delivery of captopril using hydrophilic gels has already been demonstrated [19]. The pharmacodynamics of captopril can be monitored simply via blood pressure measurements in SHRs. Here, we assess the pharmacodynamics, via blood pressure, of captopril from needle-free injection in SHRs with reference to the intravenous administration of the drug. 4.1 Materials and methods 4.1.1 Reagents The following reagents were used: captopril (Squibb), ketamine (Bioniche Animal Health Canada), xylazine and euthanyl (Biomedica-MTC Animal Health). 4.1.2 Animals In vivo experiments were conducted according to procedures previously described [20,21]. Male SHRs (10–12 weeks old, 250–350 g) were used in this study. SHRs were anesthetized with a solution of ketamine/xylazine (87/13 mg/kg) via an intraperitoneal injection with their body temperature maintained at 37 ◦ C using a heating pad. A polyethylene catheter (PE50), filled with heparin sodium (1,000 U/ml) to prevent clotting, was inserted in the right carotid artery and pushed into the aorta to continuously monitor the mean arterial blood pressure (MAP). This was measured with a pressure transducer (Micro-Med Inc. model TDX-300) connected to a blood pressure analyzer (MicroMed Inc. model BPA-100c). A second catheter (PE50) was inserted in the left jugular vein to allow bolus injection of captopril solution. At the beginning of each experiment, an average of 10–20 min of equilibration time was allowed to ensure blood pressure stabilization following catheter insertion. At the end of the experiments, animals were euthanized with a 130 μ injection of euthanyl via the carotid artery. The research protocol was approved by the Animal Research Ethics Committee at the Université de Sherbrooke. 4.1.3 Captopril injections Captopril injections were performed with the powder device. In addition, intravenous needle injections were performed for experimental validation as well as for comparison purposes.

For the powder injections, captopril powder was sieved directly from its storage container to retain particle diameters within a 20–45 μm range, which were then used to prepare pure doses for the powder injector. To simplify dose formulation and preparation, no adjuvant, excipient or filler was used. For ease of handling, two captopril powder doses in the milligram range were used, 2 and 5 mg, along with a control dose of 0 mg where no captopril was injected. Two different injector-operating conditions were used, one with an evacuated driven section (condition A) and the other with an atmospheric driven section (condition B). Needle-free injections were applied to the shaved abdomen of the animals. For all powder injection experiments, a control (0 mg) injection was first performed and measurements were recorded for 15 min. Then, either a 2 or 5 mg injection was performed on the same animal around the same site and the data subsequently recorded for 20 min. Intravenous needle injections were performed with a solution of captopril at a dose of 100 μg/kg diluted into 100 μ sterile isotonic saline (0.9%) and administered as a bolus injection via the jugular catheter. After each needle injection, the catheter was flushed with 200 μ of saline. Control injections for the needle experiments were performed with a 100 μ saline injection followed by a 200 μ saline flush. The MAPs for the needle control experiments were recorded for 5 min while needle captopril injections were recorded for between 10 and 20 min.

4.2 Results 4.2.1 Powder injections Control injections were performed with an empty drug compartment in the powder injector, i.e., a 0 mg dose, to examine animal response to the triggering of the device. The top trace of Fig. 11 shows an example for the evolution of the MAP following a powder control injection, taking place at the vertical arrow on the plot. It is observed that the MAP prior to the injection is about 98 mmHg and that the injection causes a sudden increase in MAP of about 10 mmHg. The MAP then rapidly stabilizes, within 30 s, to its pre-injection value. Although the powder injection device is not very noisy or painful, the sudden triggering of the device is probably the cause of the short increase in MAP in the animals. This sudden increase in MAP appears for experiments with the powder injector; for all these control cases, the MAP reverts to its preinjection level. Animal variability causes the time required for MAP stabilization to vary from a few seconds to a few minutes. Three different test conditions were examined with the powder injection of captopril:

123

M. Brouillette et al.

codynamic quantities useful for comparison of the various tests: Powder injector - Empty

p0 the average MAP before injection pmax the maximum MAP peak following the injection pmin the minimum MAP achieved following injection te the time between the injection and the observation of minimum pressure pmin – t ∗ the time between the observation of maximum pressure pmax and minimum pressure pmin .

– – – –

Powder injector - 2 mg captopril

Powder injector - 5 mg captopril

10 min. Intravenous needle injection. 100 ug/kg

Fig. 11 Time evolution of MAP following 0 mg, 2 mg and 5 mg captopril injection with the powder injector (condition A) and 100μg/kg captopril intravenous injection. The vertical arrow indicates t = 0, the time at which each injection took place

Normalized pressure (%)

140 120 100 80 5 mg - pressure condition A 2 mg - pressure condition A 2 mg - pressure condition B

60 40 -5

0

5

10

15

20

Time (min) Fig. 12 Time evolution of normalized MAP following captopril injection with the powder injector at three different injector-operating conditions

– a 2 mg captopril injection at injector-operating condition A – a 2 mg captopril injection at condition B – a 5 mg captopril injection at condition A Figure 11 shows examples for the time evolution of the MAP results for two of these conditions. Typically, following the initial peak in response to the injection, the MAP decreases by between 12 to 22 mmHg from its pre-injection value within 5 min, regardless of dosage and injector-operating conditions. Figure 12 shows the powder injection data in more detail, with the MAP normalized by its pre-injection value for each condition. We observe no significant difference between these three tests. Figure 13 shows on the same plot the time evolution of MAP for both control (0 mg) and 5 mg powder injections at condition A. This allows for the definition of basic pharma-

123

Table 3 shows a comparison of some of these results for the different test conditions. It is seen that, within experimental error, the results between the three test conditions are similar, with an average decrease in MAP p0 − pmin of 17±8 mmHg, a time to minimum pressure tmin of 4.2±1.2 min and a minimum pressure of 79±6 mmHg. 4.2.2 Intravenous injections The bottom trace of Fig. 11 shows an example for the evolution of the MAP following an intravenous injection of captopril. In this case, the average pre-injection pressure is around 115 mmHg; the injection, taking place at t = 0, does not provoke a sudden increase in pressure upon injection as with the powder device. This is because the needle injection takes place via a catheter, which is already in place for between 10 to 20 min prior to the injection; this therefore eliminates the “injection stress” which causes this pressure increase. Following the captopril needle injection, the MAP is seen to decrease rapidly, within 1–2 min, by about 40 mmHg. The data for all needle injections are summarized in Table 3. The average decrease in MAP is 35±12 mmHg, the time to minimum pressure is 1.6±1.3 min and the minimum pressure is 62±6 mmHg. 4.3 Discussion Every one of the 11 animals tested with the powder injector exhibited pharmacodynamic response following captopril administration. Furthermore, within experimental error, there does not appear to be differences between the results from the three operating conditions. The effect of the different injector-operating conditions for the 2 mg doses does not appear to be significant. Physically, these correspond to the same driver pressure, but different driven section initial conditions. The evacuated driven section of condition A produces higher nozzle velocities and a matched nozzle exit pressure. The atmospheric driven section of condition B produces lower nozzle velocities, but an underexpanded nozzle exit condition which results is an additional acceleration outside of the nozzle. The overall result is that particle velocities at the target are nearly the same in

A new biolistic intradermal injector

120

MAP (mmHg)

Fig. 13 Comparison between time evolution of MAP following control and captopril injection with the powder injector on the same animal—injector-operating condition A. p0 is the average MAP before injection, pmax the maximum MAP peak following the injection, pmin the minimum MAP achieved following injection, te the time between the injection and the observation of minimum pressure and t ∗ the time between the observation of maximal and minimal MAP

pmax

90

p pmin

60

t* te

30 Control (0 mg) powder injection 5 mg captopril powder injection

0 -2

0

2

4

6

8

10

12

Time (min) Table 3 Comparison between measured pharmacodynamic parameters for various test conditions IV intravenous needle injection, Powder intradermal powder delivery

Parameter

p0 − pmin (mmHg) tmin (min) pmin (mmHg)

IV

Powder 2 mg A

Powder 2 mg B

Powder 5 mg A

Powder—All tests

(n = 4)

(n = 3)

(n = 4)

(n = 4)

(n = 11)

35 ± 12

14 ± 3

18 ± 12

18 ± 7

17 ± 8

1.6 ± 1.3

4.4 ± 0.5

4.5 ± 1.2

3.7 ± 1.5

4.2 ± 1.2

62 ± 6

79 ± 4

78 ± 8

80 ± 7

79 ± 6

both cases, as validated from PIV experiments, and therefore particle penetration is identical, as shown from the captopril results. This is an important result, as the injector-operating condition B is easier to achieve in practice—the driven section is simply open to the atmosphere. The effect of dosage for the pressure condition A shows that there are no significant pharmacodynamic differences between 2 and 5 mg. Even at 2 mg, the captopril dosage corresponded to about 7 mg/kg, which greatly exceeded the intravenous dosage of 100 μg/kg and thus may have led to saturation, although the mechanisms of action of captopril have not been fully elucidated [22]. Another possibility is that the powder injector ejected the same amount of captopril in both cases, although this has not been verified. In comparison with powder injection, intravenous injections obviously resulted in faster, but also stronger response. In the case of intravenous delivery, all the injected captopril immediately reaches the blood compartment, while with intradermal delivery the drug has to diffuse into the vascular system. The latter process is slower and not all of the captopril may ultimately reach the desired target, thus explaining the faster and stronger response for intravenous delivery. Powder dosages were about a 100-fold above those for the intravenous injections; these large values were chosen for ease of handling and loading of the powder injector, since no excipient, adjuvant or filler was used to prepare the doses. The goal of the study was not to reproduce the same pharmacodynamic profile as with the intravenous injection, but to demonstrate the capability of the powder injector to deliver a

systemic drug. The effects of particle size, size distribution, velocity, shape and crystal structure, for example, were not investigated, although they may lead to different pharmacodynamic/pharmacokinetic profiles. These parameters should be optimized for each particular application. For convenience, the only injection site investigated for the powder injection experiments was on the abdomen, and no specific skin reaction was observed at this site for any of the animals tested.

5 Vaccine delivery Finally, we demonstrate the capability of the powder injector to inject a vaccine and produce the expected immunological response. In particular, we compare the delivery of an inactivated influenza vaccine in mice via subcutaneous needle injection and intradermal powder delivery. It was found that, for similar dosage, the needle-free injection device induced serum antibody response to the influenza vaccines comparable to that of subcutaneous needle injection, but without requiring the use of an adjuvant. 5.1 Materials and methods 5.1.1 Reagents The following reagents were used: HyPure Cell Culture Grade Water (HyClone), phosphate buffer saline (PBS) solu-

123

M. Brouillette et al.

tion (10×) (Sigma), trehalose dihydrate (≥99.5 %, Fluka). As adjuvant, we used TiterMax Gold (Sigma), a squalene-based oil-in-water emulsion, similar to Chiron’s (Novartis) MF59 commercial adjuvant used worldwide in seasonal influenza vaccination. 5.1.2 Vaccine The formalin-inactivated Aichi virus (A/Aichi/68 (H3N2)) was obtained from Charles River Laboratories. To prepare powder vaccines for intradermal powder delivery, 0.5 ml of the influenza vaccine solution (2 mg/ml) was incorporated into a trehalose solution prepared using 100 mg of trehalose in 20 ml of sterile water. The influenza/trehalose solution was then lyophilized in a Virtis Sentry (Model 24DX24) specimen freeze-dryer for 48 h at a pressure of 260 mTorr and a temperature of −55 ◦ C. The dried solid was collected, ground with a mortar and pestle, and sieved using stainless steel 3-in. sieves (Tyler) to collect particles between 20 and 45 μm in diameter. Each 1 mg of powder therefore contained 10 μg of influenza vaccine (total viral protein). A similar powder was formulated but without the vaccine for intradermal powder delivery control experiments. In all cases, the powder was stored in a freezer at −20 ◦ C until administration. Although the intramuscular route is commonly used for administering human vaccines, subcutaneous, intraperitoneal and intramuscular routes are often used interchangeably for immunizing experimental animals such as mice. To prepare vaccines for subcutaneous needle injection, the original vaccine solution was diluted in PBS to a concentration of 2 μg of vaccine for 10 μl of solution. This solution was then emulsified with an equal amount of adjuvant. The resulting concentration of the adjuvanted vaccine is 1 μg/10 μl. A subcutaneous needle injection control solution was prepared in a similar manner with PBS and adjuvant in equal proportions, but without vaccine. 5.1.3 Immunization BALB/c mice (female, 19–20 g; Charles River Laboratories) were used; the animals were not anesthetized during the experiments. For intradermal powder delivery, two dosages of the powder vaccine were studied, resulting in the delivery of either 5 or 10 μg (nominal) of vaccine. Intradermal powder injections were applied to the shaved abdomen of the animals. There was no bleeding, edema or any other macroscopically detectable local reactions at the vaccination sites immediately after vaccination and up to 28 days after immunization. Subcutaneous needle injection experiments were performed with the liquid using a 26.5-gauge needle—the subcutaneous injection was given in the scruff of the neck. Two

123

Table 4 Test conditions for the vaccine experiments Test condition

Number of mice

0.5 mg powder–5 μg influenza–no adjuvant

8

1 mg powder–10 μg influenza–no adjuvant

10

0.5 mg powder–0 μg influenza–no adjuvant

8

1 mg powder–0 μg influenza–no adjuvant

8

50 μl liquid–5 μg influenza–with adjuvant

9

100 μl liquid–10 μg influenza–with adjuvant

9

50-100 μl liquid–0 μg influenza–with adjuvant

4

dosages of the liquid vaccine were studied, resulting in the delivery of either 5 or 10 μg of vaccine. Control experiments were performed with the same quantities of liquid containing no vaccine. Table 4 shows a summary of the different test conditions. These dosages are consistent with other influenza vaccine studies in mice using a variety of delivery methods, e.g., [7,23]. The research protocol was approved by the Animal Research Ethics Committee at the Université de Sherbrooke. 5.2 Results Blood was drawn from the mice at 14 and 28 days and antibody responses were measured using enzyme-linked immunosorbent assay (ELISA). 5.2.1 Negative control experiments Negative, i.e., no vaccine, control experiments were performed for both powder and needle injections (see Table 4). Optical density measurements were performed on the negative control samples after ELISA at 14 days and no optical density could be observed. This demonstrated that neither the trehalose powder nor the liquid adjuvant induced any immune response. Following these results, it was decided that titers were not to be determined for these tests at both 14 and 28 days. 5.2.2 Immunization experiments Figure 14 shows the results for serum immunoglobulin (Ig) G antibody titer to influenza virus from individual mice and the mean titer of all mice receiving the same treatment, at both 14 and 28 days. It was found that, after 14 days, intradermal powder delivery of the vaccine elicited titers similar to those induced by subcutaneous injection of both 5 μg and 10 μg vaccine dosages. After 28 days, it was seen that the titer was slightly higher with the powder injector for both influenza dosages. From statistical analysis, the hypothesis that the two means are equal cannot be rejected at the 5 % significance

A new biolistic intradermal injector 6

Titer from individual mice

Mean Titer

IgG Titer (log10)

5

4

3

2

1 Titers at 14 days

Titers at 28 days

0

Titer #1: 5ug of Titer #1: 5ug of Titer #1: 10ug of Titer #1: 10ug of Titer #2: 5ug of Titer #2: 5ug of Titer #2: 10ug of Titer #2: 10ug of influenza influenza with influenza with influenza with influenza with influenza with influenza with influenza with with needle needlefree needle syringe needle syringe needlefree needlefree needlefree needle syringe syringe injector injector injector injector

Fig. 14 Comparison of the immunogenicity of formalin-inactivated Aichi/68 influenza virus administered via powder injector or subcutaneous needle. Each mouse was vaccinated once with 5 μg or 10 μg of vaccine by either administration route. Serum antibody titers on days

14 and 28 were determined by ELISA. Data represent IgG titers from individual mice () and the mean titer (−) of all mice receiving the same treatment. The arrows indicate data obtained with the powder injector

level (Student t test); the responses for both needle and powder injections can therefore be assumed as similar.

Other studies with hepatitis B [29] and rabies [30] vaccines have shown even better efficacy for intradermal needle injection as compared to the muscular route. However, intradermal needle injection is a more difficult procedure to perform. Trehalose-based powder vaccine formulation may offer better stability without refrigeration than the current liquidbased vaccines [31]. Obviously, the elementary method of vaccine powder preparation used in the present study would be replaced at large scale with cGMP spray-drying or spray freeze-drying powder formulation techniques. We are currently examining various influenza formulations with these powder preparation techniques and at the same time assessing the efficacy and stability of these various methods in mice models. Mouse models are commonly used in the early stages of influenza vaccine development, although the ferret is commonly considered as the best animal model for influenza. At this stage, for practical considerations, we will continue to use the mouse to develop the influenza powder formulation; in parallel, a human trial has been also planned when the formulation stability work will be complete. Because the powder injector works by forcing particles at high speed into the target tissue, this delivery method might be used to deliver most, if not all, vaccines now in use. Other benefits of this injector include better patient acceptance than needle injection, as well as the elimination of needle-stick injuries and the reduction of dangerous medical waste.

5.3 Discussion Of importance in these results is the fact that, at the same dosage, IgG titers are similar for both administration routes. However, the liquid vaccine used a powerful squalene-based adjuvant to elicit the same response as the powder vaccine which used no adjuvant. A previous study with squaleneadjuvanted (MF59) influenza vaccines showed that subcutaneous immunization in mice resulted in IgG titers at least an order of magnitude larger with the adjuvanted vaccine as compared to the non-adjuvanted one [24]. Similar results are available for human studies [25,26]. From these results it can thus be hypothesized that the powder delivery system is a more efficient vaccine delivery method than subcutaneous needle injection without adjuvant. The potentially higher efficacy of intradermal powder delivery is attributable to the capacity of this method to target vaccines into the immunologically potent epidermis, in close proximity to the antigen-presenting Langerhans cells. The efficacy of intradermal powder injection demonstrated in the present study is consistent with results of previous clinical studies of intradermal needle injection of influenza vaccine, where only 20 to 40 % of the dose was required as compared to an intramuscular dose to elicit an equivalent antibody response and seroconversion rate [27,28].

123

M. Brouillette et al.

As discussed above, however, this study did not involve the characterization of potential localized damage to cells and tissues due to particle penetration and/or momentary exposure to high-pressure inert gas. This damage may have a deleterious influence on the efficacy of delivery by reducing antigen uptake or slowing the dissemination of the drug to the proper target within the body. The present experiments show that delivery of a drug and of a vaccine was nonetheless successful with the new intradermal delivery device.

6 Conclusion We have presented a novel intradermal drug delivery device, exploiting the unsteady high-speed flow produced by a miniature shock tube to entrain drug or vaccine particles onto a skin target. A study of pain and physiological response of human subjects study was performed, comparing the powder injector to intramuscular needle injections. This clinical study, performed according to established pain assessment protocols, demonstrated that every single subject felt noticeably less pain with the powder injector than with the needle injection. Regarding local tolerance and skin reaction, bleeding was observed on all volunteers after needle injection, but on none of the subjects following powder injection. Measurements of physiological parameters (heart rate, blood pressure and galvanic skin response) were not significantly different for both injection methods. All volunteers expressed preference for the powder device for future injection as compared to the usual needle injection. An assessment of the pharmacodynamics, via blood pressure, of pure captopril powder from powder injection in spontaneously hypertensive rats (SHRs) was also performed. Every animal tested with the powder injector exhibited pharmacodynamic response following captopril injection. On average, results show a decrease in mean arterial pressure of 17±8 mmHg and a time to minimum pressure of 4.2±1.2 min (n = 11). Finally, we compared the delivery of an inactivated influenza vaccine in mice via subcutaneous needle injection and the powder injector. It was found that, for similar dosage, the powder device induced serum antibody response to the influenza vaccines comparable to that of subcutaneous needle injection, but without requiring the use of an adjuvant. It can therefore be inferred that, at the same dosage, the powder injection device is more efficient than subcutaneous delivery without adjuvant. Although no effort was made to optimize the formulation or the injection parameters in the present study, the powder injector demonstrates great promise for the rapid, safe and painless intradermal delivery of systemic drugs and vaccines.

123

Acknowledgments We offer our special thanks to Mr. Benoit Couture for his contribution to the design and fabrication of the device. This work is supported by the Natural Sciences and Engineering Research Council of Canada.

References 1. Klein, T.M., Wolf, E.D., Wu, R., Sanford, J.C.: High-velocity microprojectiles for delivering nucleic acids into living cells. Nature 327, 70–73 (1990) 2. Sanford, J.C., Wolf, E.D., Allen, N.K.: Method for transporting substances in to living cells and tissues and apparatus therefore. US Patent 4,945,050 (1990) 3. Bellhouse, B.J., Sarphie, D.F., Greenford, J.C.: Needless syringe using supersonic gas flow for particle delivery. US Patent 5,899,880 (1999) 4. Nakada, M., Menezes, V., Kanno, A., Hosseini, S.H.R., Takayama, K.: Shock wave based biolistic device for DNA and drug delivery. Jpn. J. Appl. Phys. 47(3), 1522–1526 (2008) 5. Nagaraja, S.R., Rakesh, S.G., Prasad, J.K., Barhai, P.K., Jagadeesh, G.: Investigations on micro-blast wave assisted metal foil forming for biomedical applications. Int. J. Mech. Sci. 61(1), 1–7 (2012) 6. Burkoth, T.L., Bellhouse, B.J., Hewson, G., Longridge, D.J., Muddle, A.G., Sarphie, D.F.: Transdermal and transmucosal powdered drug delivery. Crit. Rev. Ther. Drug 16, 331–384 (1999) 7. Chen, D.X., Endres, R.L., Erickson, C.A., Weis, K.F., McGregor, M.W., Kawaoka, Y., Payne, L.G.: Epidermal immunization by a nee- dle-free powder delivery technology: immunogenicity of influenza vaccine and protection in mice. Nat. Med. 6, 1187–1190 (2000) 8. Lesinski, G.B., Smithson, S.L., Srivastava, N., Chen, D.X., Widera, G., Westerink, J.: A DNA vaccine encoding a peptide mimic of Streptococcus pneumoniae serotype 4 capsular polysaccharide induces specific anti-carbohydrate antibodies in Balb/c mice. Vaccine 19, 1717–1726 (2001) 9. Brouillette, M.: Needleless syringe for the subcutaneous delivery of therapeutic agents. US Patent 7,320,677 (2008). 10. Brouillette, M., Dufresne, S.: Needleless syringe for the subcutaneous injection of droplets or liquid substances. US Patent 7,803,129 (2010). 11. Quinlan, N.J., Kendall, M.A.F., Bellhouse, B.J., Ainsworth, R.W.: Investigations of gas and particle dynamics in first generation needle-free drug delivery devices. Shock Waves 10, 395–404 (2001) 12. Truong, N.K., Liu, Y., Kendall, M.A.F.: Gas and particle dynamics of a contoured shock tube for pre-clinical microparticle drug delivery. Shock Waves 15, 149–164 (2006) 13. Kendall, M.A.F., Mitchell, T., Wrighton-Smith, P.: Intradermal ballistic delivery of micro-particles into excised human skin for pharmaceutical applications. J. Biomech. 37, 1733–1741 (2004) 14. Brouillette, M.: Shock waves at microscale. Shock Waves 13, 3–12 (2003) 15. Task of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability: standards of measurement, physiologic interpretation and clinical use. Circulation. 93, 1043–1065 (1996) 16. Drummond, P.D.: The effect of pain on changes in heart rate during the Valsalva manoeuvre. Clin. Auton. Res. 13, 316–320 (2003) 17. Tousignant-Laflamme, Y., Goffaux, P., Bourgault, P., Marchand, S.: Different autonomic responses to experimental pain in IBS patients and healthy controls. J Clin. Gastroenterol. 40(9), 814–820 (2006) 18. Leibold, R.A., Gilula, L.A.: Sterilization of barium for vertebroplasty: an effective, reliable, and inexpensive method to sterilize

A new biolistic intradermal injector

19.

20.

21.

22. 23.

24.

25.

powders for surgical procedures. Am. J. Roentgenol. 179(1), 198– 200 (2002) Wu, P.C., Huang, Y.B., Chang, J.J., Chang, J.S., Tsai, Y.H.: Evaluation of pharmacokinetics and pharmacodynamics of captopril from transdermal hydrophilic gels in normotensive rabbits and spontaneously hypertensive rats. Int. J. Pharm. 209, 87–94 (2000) Gobeil, F., Filteau, C., Pheng, L.H., Jukic, D., Nguyen-Le, X.K., Regoli, D.: In vitro and in vivo charaterization of bradykinin B2 receptors in the rabbit and the guinea pig. Can. J. Physiol. Pharmacol. 74, 137–144 (1996) Gobeil, F., Montagne, M., Inamura, N., Regoli, D.: Characterization of non-peptide bradykinin B2 receptor agonist (FR 190997) and antagonist (FR 173657). Immunopharmacology 43, 179–185 (1999) Katzung, B.G.: Basic and Clinical Pharmacology. USA Press, USA (1992) Garg, S., Hoelscher, M., Belser, J.A., Wang, C., Jayashankar, L., Guo, Z., Durland, R.H., Katz, J.M., Sambhara, S.: Needle-free skin patch delivery of a pandemic Influenza vaccine protects mice from lethal viral challenge. Clin. Vaccine Immunol. (2007). doi:10.1128/ CVI.00450-06 Higgins, D.A., Carlson, J.R., Van Nest, G.: MF59 adjuvant enhances the immunogenicity of influenza vaccine in both young and old mice. Vaccine 14, 478–484 (1996) Minutello, M., Senatore, F., Cecchinelli, G., Bianchi, M., Andreani, T., Podda, A., Crovari, P.: Safety and immunogenicity of an inactivated subunit influenza virus vaccine combined with MF59

26.

27.

28.

29.

30.

31.

adjuvant emulsion in elderly subjects, immunized for three consecutive influenza seasons. Vaccine 17, 99–104 (1999) Podda, A., Del Giudice, G.: MF59-adjuvanted vaccines: increased immunogenicity with an optimal safety profile. Expert Rev. Vaccines 2, 197–203 (2003) Belshe, R.B., Newman, F.K., Cannon, J., Duane, C., Treanor, J., Van Hoecke, C., Howe, B.J., Dubin, G.: Serum antibody responses after intradermal vaccination against influenza. N. Engl. J. Med. 351, 2286–2294 (2004) Kenney, R.T., Frech, S.A., Muenz, L.R., Villa, C.P., Glenn, G.M.: Dose sparing with intradermal injection of influenza vaccine. N. Engl. J. Med. 351, 2295–2301 (2004) Fadda, G., Maida, A., Masia, C., Obino, G., Romano, G., Spano, E.: Efficacy of hepatitis B immunization with reduced intradermal doses. Eur. J. Immunol. 3, 176–180 (1987) Bernard, K.W., Mallonee, J., Wright, J.C., Reid, F.L., Makintubee, S., Parker, R.A., Dwyer, D.M., Winkler, W.G.: Preexposure immunization with intradermal human diploid cell rabies vaccine. Risk and benefits of primary and booster vaccination. J. Am. Med. Assoc. 257, 1059–1063 (1987) Maa, Y.-F., Ameri, M., Shu, C., Payne, L.G., Dexiang, C.: Influenza vaccine powder formulation development: Spray-freeze-drying and stability evaluation. J. Pharm. Sci. 93, 1912–1923 (2004)

123