A novel approach to sterile pharmaceutical freeze

http://informahealthcare.com/pdt ISSN: 1083-7450 (print), 1097-9867 (electronic) Pharm Dev Technol, 2014; 19(1): 73–81 ! 2014 Informa Healthcare USA, Inc. DOI: 10.3109/10837450.2012.752388

RESEARCH ARTICLE

A novel approach to sterile pharmaceutical freeze-drying 1

The National Centre for Product Design and Development Research and 2The Cardiff School of Health Sciences, Cardiff Metropolitan University, Western Avenue, Llandaff, Cardiff, UK, and 3MicroPharm Ltd, Units F&G, Carmarthenshire, UK

Abstract

Keywords

A novel approach has been developed that enables sterile pharmaceutical products to be freeze-dried in the open laboratory without specialist facilities. The product is filled into vials, semi-stoppered and sealed inside one, followed by a second, sterilization pouch under class 100 conditions. The product is then freeze-dried in the laboratory where the vials are shelfstoppered before being returned to class 100, unwrapped and crimped. The sterilization pouches increased the resistance to water vapor movement during sublimation, thereby increasing the sublimation time and product temperature. Ovine immunoglobulins were double wrapped and lyophilized (as above) adjusting the primary drying time and shelf temperature for increased product temperature and, therefore, prevention of collapse. Ovine immunoglobulin G formulations freeze-dried to 1.1% residual moisture with no effect on protein aggregation or biological activity. The process was simulated with tryptone soya broth and no growth of contaminating microbial cells was observed after incubation at 35 C for 2 weeks. Although increasing lyophilization time, this approach offers significant plant and validation cost savings when sterile freeze-drying small numbers of vials thereby making the manufacture of treatments for neglected and orphan diseases more viable economically.

Biopharmaceutical, containment, disposables, lyophilization, sterile processing

Introduction Time to market, cost effectiveness and flexibility are key issues in today’s biopharmaceutical market and the use of disposables reduces equipment, utility and labor costs. Furthermore, costs for cleaning validation, sterilization and the turnaround of process equipment are significantly reduced or negligible1,2 when disposables are used. The advantages of disposables for bioprocesses are becoming a reality and it is common to see disposable units replacing traditional stainless steel equipment in biopharmaceutical plants3. Current research investigating the use of disposables and novel containers for freeze-drying is in its infancy4 with the most recent investigation into the characterization of syringes5. This study introduces the use of disposable containment systems for sterile pharmaceutical freeze-drying and shows that it is possible to perform sterile freeze-drying in an open laboratory environment. Using this approach there is no need to house a freeze-drier in a controlled environment, steam sterilize the freeze-drier or perform extensive validation for sterilization and cleaning. This is potentially desirable to manufacturers of orphan or neglected drugs where the need to reduce costs is critical to allow small batch size drugs to reach small patient numbers6. It is now widely recognized that rare diseases and orphan drugs provide niche opportunities for biopharmaceutical companies7 and in 2009 orphan drugs constituted 38% of the 29 new therapies that the US Food and Drug Administration (FDA) approved for marketing8.

Address for correspondence: Christopher Lee Albert Cherry, The National Centre for Product Design and Development Research, Cardiff Metropolitan University, Western Avenue, Llandaff, Cardiff CF5 2YB, UK. Tel/Fax: þ44 12 39710529. E-mail: [email protected]

History Received 30 August 2012 Revised 3 October 2012 Accepted 19 November 2012 Published online 15 January 2013

Furthermore, the global orphan drug market reached $84.9 billion in 2009 and is expected to reach $112.1 billion by 20149. The concept of sterile freeze-drying in an unclean environment has been investigated using an aluminum box incorporating a system for vial stoppering and air filtration10. The filter system allowed exchange of water vapor during freeze-drying of serum and was shown to exclude an aerosol of Serratia marcescens in challenge testing. Product temperature was monitored through thermocouple placement and showed an increase during freezedrying. More recently, W. L. Gore & Associates Inc developed the Gore Lyoguard which utilizes an ePTFE membrane allowing water vapor movement during freeze-drying. In a recent study11, this membrane was found to be able to retain 1 mm latex beads and also to prevent contamination of nutrient-rich media contained within the apparatus when challenged with microorganism. As with the aluminum box10, the Lyoguard causes an increase in product temperature during sublimation and was found to increase the total sublimation time of pure water. However, unlike the box, it is designed for bulk freeze-drying and cannot contain vials. Gassler et al.11 recognized that an additional benefit of containment in freeze-drying was the restriction of product ablation which has been shown to cause contamination of the freeze-drier12,13. Therefore, the effective use of containment during vial freeze-drying in a multiproduct pharmaceutical facility could prevent batch-to-batch contamination, expensive cleaning validation operations would not be required and turnaround times would be greatly reduced between batches. Two commonly used sterilization pouches are produced using medical grade paper and Tyvek. These inexpensive barriers allow the passage of gas molecules but not microorganisms and are routinely used to package items that are sterilized by steam or ethylene oxide gas14. Also, they are sufficiently strong and

20 13

Pharmaceutical Development and Technology Downloaded from informahealthcare.com by University of Liverpool on 01/17/14 For personal use only.

Christopher Lee Albert Cherry1, Huw Millward1, Rose Cooper2, and John Landon3

74

C. L. A. Cherry et al.

Pharm Dev Technol, 2014; 19(1): 73–81


Figure 1. Double wrapping vials in selfsealing medical grade paper pouches. The sequence shows the standard vial tray used for all experiments being wrapped in a first then second pouch to become double wrapped. The outer pouch and then the inner pouch are then opened along the chevron seal showing the double-layer structure of the pouch.

flexible to allow shelf-stoppering of contained vials with low fiber shedding15. It was postulated that pharmaceutical products could be dispensed under sterile conditions, contained within these pouches and freeze-dried in an open laboratory without adverse effects. The aim of this study was to determine the effect of containment barriers on mass and heat transfer during sublimation of ice in semi-stoppered vials, determine the effect these barriers had upon freeze-drying a model pharmaceutical protein (ovine immunoglobulin G, IgG) by evaluation of functional and structural integrity of the protein and finally simulate the process using sterile tryptone soya broth to prove asepsis.

Materials and methods Materials Schott type I borosilicate 10 mL tubular glass vials (Adelphi Healthcare Packaging, Haywards Heath, West Sussex, UK), 20 mm freeze-drying stoppers C1404 (Stelmi, Paris, France) and 152 mm square aluminum trays of 0.7 mm thickness with a capacity for 39 10 mL vials were used for all tests. Sterile water for irrigation USP (Baxter Healthcare, Compton, Berkshire, UK) was used throughout. Chevron style self-seal sterilization pouches 305 381 mm were used for all tests. SBW pouches were formed from latex impregnated medical grade paper with a laminated polyester base (Marathon Laboratory Supplies, London, UK) and SPS Medical pouches formed from Tyvek 1073B with a laminated polyester base (MET Ltd, Ashford, Kent, UK). A Gore Lyoguard R&D style container (W L Gore and Associates, Elkton, MD) with an ePTFE permeable membrane was compared to the paper and Tyvek pouches. An Edwards Super Modulyo (Edwards, Crawley, West Sussex, UK) pilot scale freeze-drier (housed in the open laboratory) was used throughout. It was fitted with calibrated RTD temperature probes and calibrated Edwards active Pirani gauges. Data were captured using a Eurotherm Chessell (Worthing, East Sussex, UK) Model 346 chart recorder. Using the RTD probes and Pirani gauges, the product temperature at the base of the vial was recorded along with chamber pressure. Pirani gauges measure the thermal conductivity of the gas (water vapor) in the chamber and capacitance manometers measure vacuum independently of gas composition. The thermal conductivity of water vapor is 60% higher than nitrogen, therefore Pirani gauges read 60% higher than capacitance manometers16. Since capacitance manometers give more accurate vacuum values, Pirani gauges were adjusted accordingly. Data were collected after 2 h sublimation for every experiment to allow steady state to be reached17. Sublimation was

then allowed to run to completion to give an indication of total sublimation time (based on when product temperature was equal to shelf temperature and a drop in Pirani pressure was observed)16. Additionally, sublimation rates were determined gravimetrically by weighing the containers using an external balance prior to sublimation and then every 2 h or until its completion. Ovine-derived polyclonal IgG specific to fluorescein isothiocynanate (FITC) was obtained from MicroPharm Ltd (Newcastle Emlyn, Carmarthenshire, UK). Ovine anti FITC was formulated in 20 mM sodium citrate buffer (pH 5.8–6.2) at 25 g/L with either 5% w/w mannitol or 5% w/w trehalose (crystalline or amorphous excipients, respectively). Sterile tryptone soya broth (EP/USP) media and 90 mm and 45 mm tryptone soya agar plates were obtained from Oxoid (Basingstoke, Hampshire, UK). Media was dispensed using a Baxa (Englewood, CO) Repeater dispensing pump and a Baxa presterilized fluid transfer tube set and fluid transfer extension set. Ice sublimation tests All sublimation tests were performed with vials containing a 3 mL fill volume (representing a 10 mm fill depth) of water. Vials were semi-stoppered and placed into the aluminum trays with each tray tightly holding 39 vials in a close hexagonal pattern. Such semistoppered vials were studied either non-contained or double wrapped (one pouch within another) inside paper or Tyvek pouches (Figure 1). The Gore Lyoguard was filled with 30 mL of water to obtain a 10 mm fill height (taken as equivalent to 10 vials of a 3 mL fill) and placed directly on the freeze-drier shelf. RTD temperature probes were positioned on the base at the center of each vial using specially adapted stoppers that held the probes in place until the product was frozen. Vials containing the RTD probes were positioned in the center vial of each tray to represent the slowest sublimating vial. A hole was drilled into the side of the Gore Lyoguard allowing insertion of an RTD probe at the base and center and then hermetically sealed with silicon mastic. The effect on sublimation of the different barriers was studied using a shelf temperature of either 0 or 10 C. The freezing stages for both shelf temperatures were identical and were: ramp to 5 C in 30 min, held for 30 min; ramp to 5 C in 30 min, held for 30 min and ramp to 40 C in 1 h, held for 2 h. Sublimation was performed at chamber pressure of 55 mTorr (shelf temperature of 0 C) and at chamber pressure of 45 mTorr (shelf temperature of 10 C) until sublimation was completed. The Gore Lyoguard was only tested at shelf temperature 0 C using the same freezing stages and chamber pressure 31 mTorr.

A novel approach to sterile pharmaceutical freeze-drying

DOI: 10.3109/10837450.2012.752388

Heat and mass transfer characterization Data derived from the sublimation experiments were used to determine figures for resistance to sublimation for each barrier and also any effects on heat transfer. Resistance was calculated by Equation (1)5,17,18


dm P0 Pc ¼ dt Rb þ Rs

ð1Þ

where dm/dt was the sublimation rate, P0 the vapor pressure of ice, Pc the chamber pressure, Rb the barrier resistance and Rs the stopper resistance. P0 was calculated from the temperature recorded for the product19, Pc was read from the chart recorder trace. Rb was calculated by subtracting Rs from the total resistance determined experimentally for barrier systems. Rs was determined experimentally testing semi-stoppered, non-contained vials. Heat and mass transfer are coupled by5,20 dQ dm ¼ DHs dt dt

ð2Þ

where DHs ¼ 2.97 KJ/g20. The vial mean heat transfer coefficient Kv is defined by5,21 dQ ¼ Kv As ðTshelf Tvial bottom Þ dt

ð3Þ

where Tshelf (Ts) was the temperature of shelf coolant, Tvial bottom the temperature recorded at the base of the vial and As the cross sectional area of the vial (4.6 cm2). From the steady state hypothesis an overall value for Kv can be calculated based on sublimation rate data20,21 Kv ¼ DHs ðdm=dtÞ As ðTshelf Tvial bottom Þ

ð4Þ

Therefore, using the data generated from the sublimation experiments it was possible to derive figures for barrier resistance from Equation (1) and then vial mean heat transfer coefficient Kv from Equation (4). Contained freeze-drying of a model protein Collapse temperatures for the IgG formulations were determined using a Lyostat3 freeze-drying microscope (BTL Ltd, Winchester, Hampshire, UK) and enabled the following freeze-drying cycles to be developed. The mannitol formulation cycle (for non-contained and for paper and Tyvek contained vials) was: ramp to 5 C, held for 30 min; ramp to 10 C in 18 min, held for 30 min; ramp to 40 C in 30 min, held for 2 h; ramp to 15 C in 24 min, held for 2 h (annealing step); ramp to 10 C in 24 min, held for 1 h; vacuum (75 mTorr) applied, ramp to 0 C in 6 min, held for 24 h; vacuum maintained and heated to 40 C in 3 h, held for 4 h, cooled to 18 C in 1 h and stoppered under vacuum when product temperature ¼ 18 C. The trehalose formulation cycle (for non-contained and paper contained vials) was: ramp to 5 C and held for 30 min; ramp to 10 C in 18 min, held for 30 min; ramp to 40 C in 30 min, held for 2 h; vacuum (80 mTorr) applied, ramp to 0 C in 30 min, held for 24 h; vacuum maintained and heated to 40 C in 6 h 42 min, held for 6 h; vacuum maintained, cooled to 18 C in 1 h and stoppered under vacuum when product temperature ¼ 18 C. The trehalose formulation cycle (for Tyvek contained vials) was identical to non-contained and paper cycles except the primary drying stage where vacuum was applied (80 mTorr), heated to 10 C in 30 min and held for 32 h. Both contained (paper and Tyvek) and non-contained vials were dried together and primary drying times were extended until

75

the temperature of the contained vial RTD was equal to the shelf temperature (accompanied by Pirani pressure drop) before secondary drying was commenced. Secondary drying times were specific for each formulation and kept constant for both contained and non-contained vials. At the end of each drying cycle, vials where stoppered by shelf compression under vacuum. Dry layer resistance of the product was calculated at 2 h using the same method for contained, water filled vials. Assessment of freeze-dried model protein The recovery of the specific IgG antibody was assessed by measurement of the quenching of fluorescein fluorescence. Samples were diluted 100-fold in 10 mM sodium borate buffer, pH 9.2, containing 1 g/L bovine gamma globulins and 1 g/L sodium azide. From this stock solution 10 serial doubling dilutions were prepared and 200 mL of each dilution was pipetted into a black 96-well microtiter plate (Greiner Bio-One Ltd, Stonehouse, Gloucestershire, UK). To each well 100 mL of 30 nM working fluorescein solution (diluted in sodium borate buffer) was added to give a final fluorescein concentration of 10 nM. The plates were incubated in the dark for 1 h at room temperature. Fluorescence intensity was measured using a BMG Labtech Polarstar fluorimeter (BMG Labtech, Offenburg, Germany) and plotted against final sample dilution. The final dilution (D, Equation (5)) that corresponded to 50% quenching of the fluorescein fluorescence was determined and at that dilution, 5 nM fluorescein is bound to the antibody. Assuming that antibody binding sites are essentially saturated with fluorescein, the concentration of bivalent IgG antibody is 2.5 nM. Using a molecular weight of 160 kDa for sheep IgG, the estimated content of specific antibody in the sample was: 2:5 109 0:16 106 D g=L

ð5Þ

where D was the dilution factor. Total protein concentration was estimated by absorbance at 280 nm on a Beckman Du Series 660 spectrophotometer (Beckman Coulter Ltd, High Wycombe, Buckinghamshire, UK) using an extinction coefficient (1 g/L, 1 cm path length) of 1.5 for ovine IgG22. Purity of IgG was assessed by size exclusion chromatography (SEC) using a GE Superose 12 HR 10/30 column23 which was equilibrated and eluted with 20 mM sodium citrate buffer, pH 5.8–6.2, in 0.9% saline at a flow rate of 0.5 mL/ min. The eluted protein was monitored by absorbance at 280 nm. Antibody turbidity was assessed using two spectrophotometric methods before freeze-drying and after reconstitution of freezedried material using sterile water. Apparent optical density was recorded at 600 nm24 and measurement of apparent optical density between 340 and 360 nm (both 1 cm path length), expressed as an average at 350 nm25 was recorded. Residual moisture content of the freeze-dried material was determined using a coulometric Karl Fischer titrator (G. R. Scientific, Ampthill, Bedfordshire, UK). Samples were reconstituted with 3 mL of Hydranal Coulomat A (Sigma Aldrich, Poole, Dorset, UK). Around 700 mL of sample was injected into the titrator and the w/w % residual moisture calculated back from the known amount of starting solid. Media simulation A small-scale freeze-drying run of 312 vials was simulated using nutrient-rich media according to FDA26 and EU27 guidelines. A total of 8 trays of 39 vials were double wrapped and autoclaved at 121 C for 30 min. Freeze-drying stoppers, control vials and paper and Tyvek pouches were autoclaved at 121 C for 15 min. Vials were then aseptically filled in a horizontal laminar flow

76



Table 1. Effect of barriers on sublimation time, sublimation rate and product temperature. Shelf temperature (Ts) 0 C Containment

Non-contained

Paper

Tyvek

Mean sublimation time (h) Sublimation rate (g/h/vial)

16 0.319 0.325 0.337 0.332 Mean ¼ 0.328 0.008 37.5 37.5 37.5 37.0 Mean ¼ 37.38 0.25

18 0.273 0.269 0.287

20 0.240 0.253 0.247

Product T ( C) Pharmaceutical Development and Technology Downloaded from informahealthcare.com by University of Liverpool on 01/17/14 For personal use only.

10 C Non-contained 20 0.245 0.241 0.258

Paper

Tyvek

25.33 0.184 0.171 0.183

28.67 0.147 0.151 0.159

Mean ¼ 0.276 0.009 23.0 22.0 20.5

Mean ¼ 0.247 0.007 16.5 15.5 15.5

Mean ¼ 0.248 0.009 38.5 39.5 40.5

Mean ¼ 0.179 0.007 26.0 27.0 28.0

Mean ¼ 0.152 0.006 21.0 21.0 22.0

Mean ¼ 21.83 1.26

Mean ¼ 15.83 0.6

Mean ¼ 39.5 1

Mean ¼ 27.0 1

Mean ¼ 21.33 0.6

cabinet in a clean room, semi-stoppered by hand using presterilized forceps, double wrapped and finally sealed. Four trays were sealed in paper pouches and four in Tyvek pouches. In addition, seven positive control vials were filled and semistoppered. Two of the paper and two of the Tyvek packages were then transferred to a Virtis (BTL Ltd, Winchester, Hampshire, UK) freeze-drier housed in the cleanroom. Two paper and two Tyvek packaged vials along with five positive control vials were transferred to the Edwards Super Modulyo Freeze-drier housed in the laboratory. The remaining two positive control vials were deliberately contaminated by skin contact. During the filling and transfer the environment was sampled for viable particles using an SAS active air sampler (Cherwell Laboratories, Bicester, Oxfordshire, UK) and for non-viable particles using a Climet C500 laser particle counter (Optical Sciences, Moulton Park, Northampton, UK). In addition, the surfaces of the laminar flow and the clean room and laboratory freeze-driers were swabbed and analyzed using a 3 M (St Paul, MN) Clean-Trace surface ATP analyzer. The vials (packaged and positive controls) were exposed for 4 h, stoppered by shelf and incubated for 2 weeks at 35 C.

Product temperature Product temperature (Tp) determined at 2 h sublimation was lowest for non-contained vials, with vials contained in Tyvek pouches having the highest (Table 1). This was consistent for Ts ¼ 0 and 10 C with product temperature marginally decreasing with shelf temperature (Figure 2). Resistance to sublimation The resistance to water vapor evolved during sublimation for different barriers was calculated using Equation (1). Barrier resistance (Rb) was determined at 2 h sublimation where noncontained vials showed the least resistance (Figure 3) and Tyvek pouches the highest resistance to sublimation. Resistances for paper and Tyvek are shown with stopper resistance (Rs) subtracted. The resistance values appear to be independent of shelf temperature, given the variability of the results. Also, the resistance shown is for a double layer of water vapor permeable material. Heat transfer (Kv)

Results Non-contained, paper and Tyvek barrier ice sublimation tests Sublimation time Containment in either paper or Tyvek pouches caused an increase in sublimation time compared with non-contained vials. Vials contained in Tyvek pouches took longer to complete sublimation than vials contained in paper pouches whether at a shelf temperature (Ts) of 0 C or at 10 C (Table 1). Thus, the sublimation times in paper were 12.5% longer and 25% longer in Tyvek when compared to non-contained vials at Ts ¼ 0 C. Vials contained in paper were 27% longer and vials contained in Tyvek were 43% longer when compared to non-contained vials at Ts 10 C. Sublimation rate Gravimetric sublimation rates determined at 2 h were normalized as a sublimation rate per vial. Values for non-contained vials were higher compared to vials contained in paper which in turn were higher than in Tyvek pouches (Table 1) at shelf temperatures (Ts) of 0 and 10 C. As expected, sublimation rates decreased with decreasing shelf temperature.

Kv is a measure of the efficiency with which the vial transfers heat from source to sink and is the sum of three contributions: Kc from direct conduction from shelf to vial, Kr from radiative heat transfer and Kg from gas conduction between the shelf and vial5,18. For different containment systems Kv was determined using Equation (4) and compared to non-contained vials (Figure 4). Non-contained vials had the lowest Kv, paper increased Kv and Tyvek had the largest observed increase in Kv. The same increase and order was observed at Ts ¼ 0 and 10 C. As would be expected, for non-contained vials, Kv is independent of shelf temperature. The effect of a change in shelf temperature is not so clear for the contained systems due to the greater variability in the results. Sublimation tests with a propriety product The Gore Lyoguard (Lyo R&D container) was tested in direct contact with the freeze-drier shelf without the use of an aluminum tray. Sublimation rate, product temperature, resistance and heat transfer was determined at Ts ¼ 0 C only (Table 2) The Lyoguard results compared closely to non-contained vials with sublimation rates lower and product temperature 1 C higher (Table 1), double the resistance to Rs (Figure 3) and Kv


DOI: 10.3109/10837450.2012.752388


Figure 2. Effect of barriers on product temperature (Tp) showing a product temperature increasing with different barriers.

Figure 3. Effect of barriers on resistance to sublimation showing resistance increasing with different barriers.

Figure 4. Effect of barriers on vial mean heat transfer coefficient (Kv) showing Kv increasing with different barriers.

Table 2. Characteristics for the Gore Lyoguard. Product T ( C) 37.0 36.0 36.0 Mean ¼ 36.33 0.6

Sublimation rate (g/h/vial)

Resistance (Torr h/g/vial)

Heat transfer (KJ/h/cm2/ C)

0.268 0.273 0.268 Mean ¼ 0.271 0.002

0.406 0.436 0.444 Mean ¼ 0.429 0.02

4.61 103 4.81 103 4.74 103 Mean ¼ 4.72 103 1 104

77

78



Table 3. Comparison of paper and Tyvek pouches to non-contained vials showing the effects of containment on freeze-drying cycle parameters for amorphous and crystalline IgG formulations. Mannitol Containment Primary drying time (h) Secondary drying time (h) Shelf temperature ( C) Product temperature ( C)

Trehalose

Non-contained

Paper

Tyvek

Non-contained

Paper

Tyvek

18.5 10 0 32

20 10 0 23

21 10 0 16.5

20 16 0 30

22 16 0 22

31 16 10 20.5


Table 4. Comparison of sublimation rate and dry layer resistance (during primary drying) of crystalline and amorphous formulations of IgG. Mannitol Sublimation rate (g/h/vial) Resistance (Torr h/g/vial)

0.251 0.50

0.233 0.671

Trehalose 0.231 0.566

0.128 1.018

0.134 1.13

0.132 0.817

lower (Figure 4). However, Kv cannot be compared to noncontained vials as the Lyoguard was in direct contact with the shelf.

The degree of turbidity increase was comparable for all samples, contained or non-contained, and observed to be an effect of the freeze-drying process, not of containment.

Effect of containment on IgG formulations

Media simulation

Freeze-drying cycles

The two vials filled with TSB media and deliberately contaminated as positive controls became turbid, due to microbial growth, after 24 h incubation at 35 C. Of the five positive control vials for the laboratory freeze-drier, one vial showed growth after 48 h incubation. After 2 weeks the eight packages of vials were opened and each individual vial was compared to a positive control showing growth. The vials in the packages all remained clear and were therefore negative for microbial contamination. Environmental monitoring data showed that the laminar flow maintained an EU grade A environment and the clean room an EU grade B environment for both viable and non-viable particles. Swabs analyzed with the Clean-Trace (3 M) give results in relative light units (RLU) which is proportional to the amount of ATP present, and therefore gives a measure of cleanliness of a surface28; the higher the RLU, the more contaminated is a surface. Swabs of the laminar flow and Virtis freeze-drier (housed in the cleanroom) gave between 10 and 20 RLU. These figures are low and can be considered as background counts. However, the Edwards freeze-drier (housed in the laboratory) gave counts of 101–320 RLU. Although these counts are relatively low, compared to a heavily contaminated area, they still show that the laboratory drier is less clean than the cleanroom drier and significant potential for contamination of a sterile product exists.

Freeze-drying microscopy was used to determine the collapse temperatures (Tc) of amorphous (trehalose) and crystalline (mannitol) formulations of IgG. Tc was 20 C for the former and 7 C for the latter. The sublimation time for IgG was monitored and extended until completion. Ts and, therefore, Tp was carefully monitored to ensure Tp stayed several degrees below Tc. The effect of containment systems on the freeze-drying behavior of the antibody reflected that observed for pure water (Table 3). Primary drying time and product temperature increased with increased resistance to water vapor passage of each barrier used. The same order of barrier resistance was found as for ice sublimation tests. It is important to note that when using the Tyvek pouches with the trehalose formulation, the shelf temperature was decreased to 10 C. This was done to avoid primary drying above the collapse temperature. However, reduction of Ts to 10 C gave a Tp dangerously close to Tc although the drying proceeded without collapse. Dry layer sublimation rates and resistances To establish an overall comparison of resistance to water vapor during sublimation, the dry layer resistance of the amorphous and crystalline formulations were determined using the described methods (Table 4). Comparison of Table 4 to Table 2 and Figure 3 shows that dry layer resistance is indeed significant and is greater than the resistance offered by a stopper, and interestingly also the Lyoguard ePTFE membrane. However, dry layer resistance is significantly less than that of the paper or Tyvek barriers. Effects of contained freeze-drying on a model protein (ovine IgG) Fluorescence quenching showed that the antibodies in noncontained freeze-dried vials (Table 5), vials double wrapped in paper pouches (Table 6) and vials double wrapped in Tyvek pouches (Table 7) had no significant loss of biological activity and remained 495% pure with no significant loss of total protein following freeze-drying. Residual moisture for both formulations was 1.1%. Some increase in turbidity was noted upon reconstitution, significant only for the mannitol formulation with a minor increase in trehalose turbidity compared to pre-drying.

Discussion Containment increases primary drying times by affecting the mass transfer characteristics of sublimation. Containing semi-stoppered vials decreases the rate of mass transfer between the aqueous product and the chamber/condenser. This slows the overall sublimation time impacting on the product temperature and associated vapor pressure of ice within the containers29. The sublimation process uses energy through the latent heat of sublimation. To balance this energy demand, heat energy is input to the system via shelves. This stops the product cooling and reaching equilibrium with the condenser where sublimation would cease30. The reduced sublimation rates induced by the barriers reduced the heat loss through sublimation but the heat input remains the same through constant Ts. This resulted in a significant increase in product temperature for the contained vials. The degree of barrier resistance for the contained systems reflects the increase in product temperature. Resistances appear to


DOI: 10.3109/10837450.2012.752388

79

Table 5. Effects of freeze-drying on IgG characteristics in non-contained vials. Trehalose


Reconstitution time Turbidity (600 nm) Turbidity (350 nm) OD280 (g/L) SEC (% pure) Sp AB (fluorescence quenching) (g/L) Residual moisture (Karl Fischer)

Mannitol

Pre-freeze-drying

Post-freeze-drying

Pre-freeze-drying

Post-freeze-drying

N/A 0.022 0.139 26.8 495 4.5 N/A

530 s 0.058 0.213 26.1 495 4.5 1.11%

N/A 0.012 0.107 27.1 495 4.4 N/A

530 s 0.107 0.361 26.2 495 4.1 0.64%

Table 6. Effects of freeze-drying on IgG characteristics in vials semi-stoppered and double wrapped in paper pouches. Trehalose


Mannitol

Pre-freeze-drying

Post-freeze-drying

Pre-freeze-drying

Post-freeze-drying

N/A 0.022 0.139 26.8 495 4.5 N/A

530 s 0.039 0.182 26.0 495 4.3 0.81%

N/A 0.012 0.107 26.1 495 4.4 N/A

530 s 0.102 0.318 25.9 495 3.7 0.65%

Table 7. Effects of freeze-drying on IgG characteristics on vials were semi-stoppered and double wrapped in Tyvek pouches. Trehalose


Mannitol

Pre-freeze-drying

Post-freeze-drying

Pre-freeze-drying

Post-freeze-drying

N/A 0.0298 0.1595 19.84 495 4.01 N/A

530 s 0.0407 0.194 26.99 495 4.1 0.7%

N/A 0.022 0.1433 27.38 495 3.84 N/A

530 s 0.1013 0.3309 26.59 495 3.81 0.68%

be independent of shelf temperature and broadly constant. Stoppers posed the least resistance and had the shortest sublimation time and highest rate. Tyvek has the longest sublimation time and most reduced sublimation rate with paper between these two. The use of the paper/Tyvek barriers would clearly have an impact on the heat transfer characteristics of the contained vials and an interesting effect on Kv was observed. The mean heat transfer coefficients for the barrier systems increased relative to the non-contained vials. Using the same trays and vials, it would be expected that Kv would remain the same with constant Ts. As Pikal et al.18 demonstrated, vial heat transfer coefficients increase with increasing container/chamber pressure. It would appear that resistance posed by the double layer of barrier is causing an increase in pressure inside the pouch causing the Kv to increase. This pressure-dependent increase in Kv in the pouch may suggest that the dominant contribution to Kv would be from Kg as Kc and Kr are both independent of pressure5,18. However, minor changes in geometry and set up can impact Kv. For example, no attempt was made in this study to insulate the front of the drier with aluminum foil to limit Kr, which in turn would affect Kc and Kg31. Furthermore, an extra double layer, attributable to the polymer film of the pouch, was added between the shelf and the base of the aluminum tray creating yet another heat transfer interface along with the double layer of water vapor permeable barrier.

Therefore, further work would be required to determine exactly the heat transfer mechanisms at work with the barriers studied. The literature shows that manometric temperature measurement (MTM) would be a more advanced technique to calculate these data32–35, unfortunately this technology was not available for this study. Future investigation might involve MTM along with a more detailed analysis to determine the contributing factors of heat transfer. The ePTFE membrane of the Gore Lyoguard functioned effectively as a water vapor permeable barrier during sublimation and the sublimation rates were comparable to non-contained vials of which resistance has been taken as negligible in previous studies33. The high sublimation rate was also accompanied by a low product temperature, indicating a low resistance to water vapor flow. Although the figures measured for heat transfer cannot be compared directly to non-contained vials, paper or Tyvek pouches (due to direct shelf contact), the data for sublimation rate, Tp and resistance indicate that a comparable Kv would be low. Other researchers have reported that dry layer resistance is greater than the resistance offered by a stopper36,37 and to date, dry layer resistance has been the highest resistance determined during sublimation. Data from this study show that dry layer resistance is indeed significant and interestingly also greater than the Lyoguard resistance (also demonstrated by Gassler et al.)12.


80


However, dry layer resistance is significantly less than that of the paper or Tyvek barriers, where it must be remembered that a double layer has been used throughout. Hence, it is important to highlight that freeze-drying is possible using barriers that pose a greater resistance to water vapor flow than product dry layer. The effect of containment systems on the freeze-drying behavior of the antibody reflected that observed with pure water. Primary drying time and product temperature increased with increased resistance to water vapor passage of each barrier used and the same ranking of resistance is found as for ice sublimation tests. Using amorphous and crystalline formulations of the model pharmaceutical protein, primary drying times were increased and shelf temperature was adjusted to prevent collapse. The resulting products, when compared with non-contained freeze-dried IgG, showed no effect on antibody activity or structure. Residual moisture levels were also comparable, indicating that containment did not prevent bound water removal during secondary drying. Furthermore, it could also be inferred that no structural changes occurred to the freeze-dried cake as reconstitution times are the same. However, scanning electron microscopy of the cakes is required to prove this deduction. Each of the barriers were selected because of their known ability to prevent the passage of microorganisms (LRV of 5.2 for Tyvek 1073B)15, and this has been proven during a process simulation of a small batch size vial freeze-drying process. The process simulation qualifies the ability of these packages to prevent contamination of nutrient-rich media and was performed following guidelines set down by both European and US regulators. Double wrapping aseptic packages is a common practice and two layers provide extra security for the sterile contents. The use of the medical grade paper and Tyvek pouches would mean an increase in manual handling, and therefore higher potential risk for operator contamination of sterile products. However, experienced and skilled operators would be able to perform these aseptic manipulations without contaminating product, as demonstrated by this media simulation. Major advances have been achieved in the sterile freeze-drying of pharmaceutical products designed for systemic administration. Large freeze-driers installed in class 100 cleanrooms permit the freeze-drying of many thousands of vials at a time. The introduction of robotic equipment has minimized the need for operator intervention and, therefore, contamination, and there is no doubt that this approach will remain the norm for the freezedrying of products with large sales, such as monoclonal antibodies. Such measures are less suitable for orphan or niche products when only a few hundred vials require freeze-drying. It also poses problems for small companies that lack the financial resources to establish such facilities, especially for the freeze-drying of niche products for developing countries, such as antivenoms for Africa, where costs must be kept to a minimum. Much of the present cost for small-scale freeze-drying comes from the need to install the instrument in sterile facilities, ensuring sterile transfer of the vials between the filling unit and the freeze-drier, and the need for ensuring scrupulous cleanliness throughout and between each freeze-drying operation. Finally, it should be noted that all present approaches are totally unsuitable for the freeze-drying of potential pathogens or their toxins or other large toxic molecules since the immediate environment, including the freeze-drier and cleanroom will be contaminated, the described containment systems offer a solution to this. Most pharmaceutical companies, however small, will have the facilities necessary to sterile fill products intended for systemic use under cGMP. Provided that the product is protected from contamination and vice versa, it can be transferred from such facilities into a non-sterile environment. Hence the interest in the


Gore Lyoguard, which is the only commercially available disposable apparatus for contained bulk freeze-drying. Incidentally, there appears no reason why stainless steel trays currently used for bulk freeze-drying cannot be contained within the pouches described as a cost effective competitor to the Lyoguard. When comparing the Gore Lyoguard to medical grade paper and Tyvek pouches, the Lyoguard exhibited minimal resistance to water vapor. However, the minimal resistance to the passage of water vapor was due to the use of a relatively complex and expensive membrane. Having been developed as a single use bulk freeze-drying apparatus, the Lyoguard containment system does not allow the stoppering of vials upon completion of the freezedrying process. When selecting disposable equipment minimum expenditure is desired, the Lyoguard is several orders of magnitude more expensive than pouches, and therefore pouches would appear a more appropriate choice for disposable packaging for sterile freeze-drying of pharmaceutical products.

Conclusions Freeze-drying of sterile pharmaceutical products is possible in a non-sterile environment using relatively inexpensive containment systems. This potentially removes the need for steam sterilization of the freeze-drier and validation of the sterilization and cleaning processes. This represents a significant cost saving in the processing of small batches of pharmaceuticals, such as orphan and neglected drugs, and easily justifies the slightly longer primary drying times that containment causes. However, adoption of this technique would require careful consideration of the balance between process economics and regulatory compliance and could present other validation requirements that would require further investigation. Further work is being undertaken to investigate the ability of the described pouches to contain microorganism during preservation by freeze-drying. If successful, this application would assist laboratories preserving biological standards or microbial seed lots where ablation of microorganism during drying represents a serious contamination issue.

Acknowledgements This research is part of a Knowledge Economy Skills Scholarship (KESS) funded by the European Social Fund. In addition, this research has been made possible with the help and assistance of MicroPharm Ltd and BTL Ltd.

Declaration of interest The authors report no conflicts of interest.

References 1. Novais JL, Titchener-Hooker NJ, Hoare M. Economic comparison between conventional and disposables-based technology for the production of biopharmaceuticals. Biotechnol Bioeng 2001;75:143–53. 2. Rao G, Moreira A, Brorson K. Disposable bioprocessing: the future has arrived. Biotechnol Bioeng 2009;102:348–56. 3. Farid SS. Process economics of industrial monoclonal antibody manufacture. J Chromatogr B 2007;848:8–18. 4. Patel SM, Pikal MJ. Emerging freeze-drying process development and scale-up issues. AAPS PharmSciTech 2011;12:372–8. 5. Patel SM, Pikal MJ. Freeze-drying in novel container systems: characterization of heat and mass transfer in glass syringes. J Pharm Sci 2010;99:3188–204. 6. Fehr A, Thurmann P, Razum O. Drug development for neglected diseases: a public health challenge. Trop Med Int Health 2006;9:1335–8.


DOI: 10.3109/10837450.2012.752388

7. Melnikova I. Rare diseases and orphan drugs. Nature Rev Drug Discov 2012;11:267–8. 8. Cote TR, Xu K, Parsier AR. Accelerating orphan drug development. Nature Rev Drug Discov 2010;9:901–2. 9. Sharma A, Jacob A, Tandon M, Kumar D. Orphan drug: development trends and strategies. J Pharm Bioallied Sci 2010;2: 290–9. 10. Taylor R, Boardman CFB, Wallis RG. Sterile freeze-drying in an unclean environment. J Appl Chem Biotechnol 1978;28:213–16. 11. Gassler M, Rey L. Development of a new concept for bulk freezedrying: lyoguard freeze-dry packaging. In: Rey L, May JC, eds. Freeze-drying/lyophilization of pharmaceutical and biological products. Chapter 12. London: Informa Healthcare; 2004:325–48. 12. Barbaree JM, Sanchez A. Cross contamination during lyophilization. Cryobiology 1982;19:443–7. 13. Adams GDJ. The loss of substrate from a vial during freeze-drying using Escherichia coli as a trace organism. J Chem Tech Biotechnol 1991;52:511–8. 14. Nolan PJ. Sterile medical device package development. In: Kutz M, ed. Standard handbook of biomedical engineering and design. Chapter 7. New York: McGraw-Hill; 2009:181–224. 15. DuPont. DuPont medical packaging guide technical reference guide. Wilmington, DE: DuPont; 2009. 16. Patel SM, Pikal M. Process analytical technologies (PAT) in freezedrying of parenteral products. Pharm Dev Technol 2009;14:567–87. 17. Pikal MJ. Use of laboratory data in freeze-drying process design: heat and mass transfer coefficients and the computer simulation of freeze-drying. J Parenter Sci Technol 1985;39:115–38. 18. Pikal MJ, Roy ML, Shah S. Mass and heat transfer in vial freezedrying of pharmaceuticals: the role of the vial. J Pharm Sci 1984;73:1224–37. 19. Wagner W, Saul A, Pruss A. International equations for the pressure along the melting and along the sublimation curve of ordinary water substance. J. Phys Chem Ref Data 1994;23:515. 20. Searles JA. Optimizing the throughput of freeze-dryers within a constrained design space. In: Rey L, May JC, eds. Freeze-drying/ lyophilization of pharmaceutical and biological products. Chapter 18. London: Informa Healthcare; 2010:425–40. 21. Hottot A, Vessot S, Andrieu J. Determination of mass and heat transfer parameters during freeze-drying cycles of pharmaceutical products. PDA J Pharm Sci Technol 2005;59:138–53. 22. Curd J, Smith TW, Jaton JC, Haber E. The isolation of digoxin specific antibody and its use in reversing the effects of digoxin. Proc Natl Acad Sci USA 1971;68:2401–6. 23. Rawat S, Laing G, Smith DC, et al. A new antivenom to treat Eastern coral snake (Micrurus fulvius fulvius) envenoming. Toxicon 1994;32:185–90.


81

24. Al-Abdulla I, Garnvwa JM, Rawat S, et al. Formulation of a liquid ovine Fab-based antivenom for the treatment envenomation by the Nigerian carpet viper (Echis Ocellatus). Toxicon 2003;42:399–404. 25. Eckhardt BM, Oeswein JQ, Yeung DA, et al. A turbidimetric method to determine visual appearance of protein solutions. J Pharm Sci Technol, 1994;48:64–70. 26. US FDA. Lyophilization of parenterals (7/93), guide to inspections of lyophilization of parenterals. Silver Spring (MD): US FDA; 2009. 27. European Commission. EU guidelines to good manufacturing practice medicinal products for human and veterinary use. Vol. 4. Annex 1, manufacture of sterile medicinal products, 2010. Available from: http://ec.europa.eu/health/files/eudralex/vol-4/2008_11_25_ gmp-an1_en.pdf [last accessed 31 Dec 2012]. 28. Corbitt AJ, Bennion N, Forsythe SJ. Adenylate kinase amplification of ATP bioluminescence for hygiene monitoring in the food and beverage industry. Lett App Microbiol 2000;30:443–7. 29. Tang X, Pikal MJ. Design of freeze-drying processes for pharmaceuticals: practical advice. Pharm Res 2004;21:191–200. 30. Adams GDJ. Freeze-drying of biological materials. Drying Technol 1991;9:891–925. 31. Rambthatla S, Pikal MJ. Heat and mass transfer scale up issues during freeze-drying, I: atypical radiation and edge vial effect. AAPS PharmSciTech 2003;4:E14. 32. Tang X, Nail SL, Pikal MJ. Evaluation of manometric temperature measurement, a process analytical tool for freeze-drying: part I. Product temperature measurement. AAPS PharmSciTech 2006;7:14. 33. Tang X, Nail SL, Pikal MJ. Evaluation of manometric temperature measurement, a process analytical tool for freeze-drying: part II measurement of dry layer resistance. AAPS PharmSciTech 2006;7:93. 34. Tang X, Nail SL, Pikal MJ. Evaluation of manometric temperature measurement, a process analytical tool for freeze-drying: part III: heat and mass transfer measurement. AAPS PharmSciTech 2006;7:97. 35. Tang X, Nail SL, Pikal MJ. Freeze-drying process design by manometric temperature measurement: design of a smart freezedryer. Pharm Res 2005;22:685–700. 36. Kuu WY, Hardwick LM, Akers M. Rapid determination of dry layer mass transfer resistance for various pharmaceutical formulations during primary drying using product temperature profiles. Int J Pharm 2006;313:99–113. 37. Johnson RE, Oldroyd ME, Ahmed SS, et al. Use of manometric temperature measurements (MTM) to characterize the freeze-drying behavior of amorphous protein formulations. J Pharm Sci 2010; 99:2863–73.