Efficient use of longitudinal CD4 counts and viral load measures in survival analysis.

CD4 counts and viral loads are dynamic quantities that change with time in HIV-infected persons. Commonly used single summary measures, such as viral load set point or early CD4 count, do not explicitly account for changes in viral load or CD4 counts or other features of the overall time course of these measures. However, the efficient use of all repeated measurements within each subject is often a challenge made more difficult by sparse and irregular sampling over time. Here, we illustrate how functional principal component (FPC) analysis provides an effective statistical approach for exploiting the patterns in CD4 count and viral load data over time. We demonstrate the method by using data from Kenyan women who acquired HIV-1 during follow-up in a cohort that practices high-risk activities and were subsequently followed up prospectively from early infection. The FPC scores for each woman obtained using this method served as informative summary statistics for the CD4 count and viral load trajectories. Similar to baseline CD4 count or viral set point, the first FPC score can be interpreted as a single-value summary measure of an individual's overall CD4 count or viral load. However, unlike most single-value summaries of CD4 count or viral load trajectories, the first FPC score summarizes the dynamics of these quantities and is seen to reveal specific features of the trajectories associated with mortality in this cohort. Moreover, the FPC scores are shown to be a more powerful prognostic factor than other common summaries when used in survival analysis.

Enzymatic oxidation of cholesterol aggregates in supercritical carbon dioxide.

Fundamental studies of enzyme-solvent interactions can be conducted with supercritical fluids because small changes in pressure or temperature may bring about great changes in the properties of a single solvent near its critical point. Cholesterol oxidase is active in supercritical carbon dioxide and supercritical carbon dioxide-cosolvent mixtures. Variations in solvent power caused by pressure changes or by the addition of dopants affected the rate of enzymatic oxidation of cholesterol by altering the structure of cholesterol aggregates.

Supercritical fluid extractions in biotechnology.

Over the past decade, supercritical fluid (SCF) extraction has been shown to deserve consideration as an alternative to liquid-liquid extraction or distillation. Most current commercial applications of SCF extraction involve biologically produced materials; the technique may be particularly relevant to extraction of biological compounds in cases where there is a requirement for low-temperature processing, high mass-transfer rates and negligible carry over of solvent into the final product. New advances, in which extraction is combined with reaction or crystallization steps, may further increase the attractiveness of SCFs in the bioprocessing industries.

Stability of subtilisin and lysozyme under high hydrostatic pressure.

The stabilities of subtilisin and lysozyme under hydrostatic pressures up to 200 MPa were investigated for up to 7 days at 25 degrees C. Methods were chosen to assess changes in tertiary and secondary protein structure as well as aggregation state. Tertiary structure was monitored in situ with second derivative UV spectroscopy and after pressure treatment by dynamic light scattering and second derivative UV spectroscopy. Secondary structure and potential secondary structural changes were characterized by second derivative FTIR spectroscopy. Changes in aggregation state were assessed using dynamic light scattering. Additionally, protein concentration balances were carried out to detect any loss of protein as a function of pressure. For the conditions tested, neither protein shows measurable changes in tertiary or secondary structure or signs of aggregation. Lysozyme concentration balances show no dependence on pressure. Subtilisin concentration balances at high protein concentration (4 mg/mL and higher) do not show pressure dependence. However, the concentration balances carried out at 0.4 mg/mL show a clear sign of pressure dependence. These results may be explained by protein interaction with the vial surface and appear to be rate limited by the equilibrium between active and inactive protein on the surface. Pressure increases protein loss, and the estimated partial molar volume change between the two states is estimated to be -20 +/- 10 mL/mol.

Sub-micrometer-sized biodegradable particles of poly(L-lactic acid) via the gas antisolvent spray precipitation process.

Sub-micrometer-sized particles of poly(L-lactic acid) may be formed by using near-critical or supercritical carbon dioxide as an antisolvent to precipitate poly(L-lactic acid) from droplets of methylene chloride solution sprayed into a carbon dioxide continuous phase. Particle sizes may be controlled by varying the density of the carbon dioxide; at constant temperature in the supercritical region, higher carbon dioxide densities yield larger particles. Two methods (one batch and one continuous) for introducing the poly(L-lactic acid) solutions into carbon dioxide are demonstrated. Although the two methods use very different mechanisms for forming the droplets, similar particle sizes are observed as a function of carbon dioxide density. We suggest that mass transport, rather than jet breakup and hydrodynamics, controls particle sizes in the near-critical and supercritical regions.

Manipulation of lyophilization-induced phase separation: implications for pharmaceutical proteins.

Lyophilization, or freeze-drying, of pharmaceutical proteins is often the only processing method that provides requisite long-term product stability. Freezing and drying, however, can cause acute damage to proteins. To alleviate damage, formulations frequently include protein stabilizers (often polymers and/or sugars), as well as buffering salts and "inert" bulking agents. While great efforts are placed on developing a formulation and suitable lyophilization cycle, incompatibilities among components through freezing and drying have been almost completely ignored. We demonstrate that solutions of poly(ethylene glycol) (PEG) and dextran, initially below critical concentrations for phase separation, do indeed experience a liquid-liquid phase separation induced by freeze concentration during the lyophilization cycle. The separation is shown to evolve with annealing at -7 degrees C and can be effectively inhibited simply by replacing NaCl with KCl in the formulation buffer. In addition, we show that phase separation causes unfolding of a model protein, recombinant hemoglobin, when freeze-dried in the PEG/dextran system. When the phase separation is averted by switching to KCl, the protein structural damage is also avoided. Measurements of pH in the frozen solutions show that the structural damage is not a result of pH changes. We suggest that KCl forms a glass with rapid cooling which kinetically prevents the phase separation and thus the protein structural damage.

Molten globule intermediate of recombinant human growth hormone: stabilization with surfactants.

We demonstrate that a surfactant-stabilized molten globule intermediate exists for recombinant human growth hormone (rhGH), is very hydrophobic, and tends to form aggregates. Characterization of this intermediate included equilibrium denaturation measured by electron paramagnetic resonance (EPR) and CD spectroscopy, assessment of aggregation during refolding, and fluorescence studies of its binding to the hydrophobic probe, 1-anilinonapthalene-8-sulfonate (1,8-ANS). We have found that at 4.5 M guanidinium hydrochloride (GuHCl), a molten globule intermediate of rhGH is stabilized and results in significant aggregation upon refolding. This intermediate is populated by the addition of the nonionic surfactant, Tween. This surfactant also reduces the extent of aggregation during refolding of rhGH from 4.5 M GuHCl. Overall, our studies reveal that rhGH forms a molten globule-like intermediate during folding and this intermediate self-associates. This self-association is reduced upon formation of a Tween-rhGH complex. Tween also binds to the native protein. Thus, nonionic surfactants such as Tween may act like molecular chaperones in facilitating protein folding while not altering the native conformation.

Effects of phase separating systems on lyophilized hemoglobin.

Polymer liquid-liquid two-phase systems offer a unique opportunity to study the mechanisms of protein stabilization during freezing and freeze-drying. Fourier transform infrared spectroscopy was used to monitor the structural integrity of recombinant hemoglobin frozen and lyophilized in the separated phases of a polyethylene glycol (PEG)-dextran system. Protein in each phase of an equilibrated biphasic PEG-dextran system experiences similar levels of structural protection against freezing stresses despite large differences in polymer concentration. This result further demonstrates previous suggestions that proteins are protected during freezing by the preferential exclusion mechanism. There are, however, distinct differences in the level of structural protection that polymers in equilibrium phases provide to proteins during lyophilization, emphasizing that the mechanisms of protein protection during freezing and drying are fundamentally different. In addition, we provide evidence that phase separation per se occurring during the course of the lyophilization cycle can be detrimental to the structural stability of a protein.

Conformational stability of lyophilized PEGylated proteins in a phase-separating system.

PEGylation of proteins is of great interest to the pharmaceutical industry as covalent attachment of poly(ethylene glycol) (PEG) molecules can increase protein sera half-lives and reduce antigenicity. Not surprisingly, PEGylation significantly alters the surface characteristics of a protein, and consequently, its conformational stability during freezing and drying. Freeze concentration-induced phase separation between excipients has been previously shown to cause degradation of the secondary structure in lyophilized hemoglobin. In this report we show how PEGylation of two proteins, hemoglobin- and brain-derived neurotrophic factor (BDNF), influences partitioning and protein secondary structure as determined by FTIR spectroscopy in a system prone to freezing-induced phase separation. PEGylation of hemoglobin reduces the loss of structure induced by lyophilization in a PEG/dextran system that phase separates during freezing, perhaps due to altered partitioning. The partition coefficient for native hemoglobin favors the dextran-rich phase (PEG/dextran partition coefficient = 0.3), while PEGylated hemoglobin favors the PEG phase (partition coefficient = 3.1). In addition, we demonstrate that PEGylation alters hemoglobin's stability during lyophilization in the absence of other excipients. In contrast, because native BDNF already partitions into the PEG-rich phase, PEGylation of BDNF has a less dramatic effect on both partition coefficients and conformational stability during lyophilization. This is the first report on the effects of PEGylation on protein structural stability during lyophilization and points out the need to consider modification of formulations in response to changing protein surface characteristics.

The ice nucleation temperature determines the primary drying rate of lyophilization for samples frozen on a temperature-controlled shelf.

The objective of this study was to determine the influence of ice nucleation temperature on the primary drying rate during lyophilization for samples in vials that were frozen on a lyophilizer shelf. Aqueous solutions of 10% (w/v) hydroxyethyl starch were frozen in vials with externally mounted thermocouples and then partially lyophilized to determine the primary drying rate. Low- and high-particulate-containing samples, ice-nucleating additives silver iodide and Pseudomonas syringae, and other methods were used to obtain a wide range of nucleation temperatures. In cases where the supercooling exceeded 5 degrees C, freezing took place in the following three steps: (1) primary nucleation, (2) secondary nucleation encompassing the entire liquid volume, and (3) final solidification. The primary drying rate was dependent on the ice nucleation temperature, which is stochastic in nature but is affected by particulate content and the presence of ice nucleators. Sample cooling rates of 0.05 to 1 degrees C/min had no effect on nucleation temperatures and drying rate. We found that the ice nucleation temperature is the primary determinant of the primary drying rate. However, the nucleation temperature is not under direct control, and its stochastic nature and sensitivity to difficult-to-control parameters result in drying rate heterogeneity. Nucleation temperature heterogeneity may also result in variation in other morphology-related parameters such as surface area and secondary drying rate. Overall, these results document that factors such as particulate content and vial condition, which influence ice nucleation temperature, must be carefully controlled to avoid, for example, lot-to-lot variability during cGMP production. In addition, if these factors are not controlled and/or are inadvertently changed during process development and scaleup, a lyophilization cycle that was successful on the research scale may fail during large-scale production.

Annealing to optimize the primary drying rate, reduce freezing-induced drying rate heterogeneity, and determine T(g)' in pharmaceutical lyophilization.

In a companion paper we show that the freezing of samples in vials by shelf-ramp freezing results in significant primary drying rate heterogeneity because of a dependence of the ice crystal size on the nucleation temperature during freezing.1 The purpose of this study was to test the hypothesis that post-freezing annealing, in which the product is held at a predetermined temperature for a specified duration, can reduce freezing-induced heterogeneity in sublimation rates. In addition, we test the impact of annealing on primary drying rates. Finally, we use the kinetics of relaxations during annealing to provide a simple measurement of T(g)', the glass transition temperature of the maximally freeze-concentrated amorphous phase, under conditions and time scales most appropriate for industrial lyophilization cycles. Aqueous solutions of hydroxyethyl starch (HES), sucrose, and HES:sucrose were either frozen by placement on a shelf while the temperature was reduced ("shelf-ramp frozen") or by immersion into liquid nitrogen. Samples were then annealed for various durations over a range of temperatures and partially lyophilized to determine the primary drying rate. The morphology of fully dried liquid nitrogen-frozen samples was examined using scanning electron microscopy. Annealing reduced primary drying rate heterogeneity for shelf-ramp frozen samples, and resulted in up to 3.5-fold increases in the primary drying rate. These effects were due to increased ice crystal sizes, simplified amorphous structures, and larger and more numerous holes on the cake surface of annealed samples. Annealed HES samples dissolved slightly faster than their unannealed counterparts. Annealing below T(g)' did not result in increased drying rates. We present a simple new annealing-lyophilization method of T(g)' determination that exploits this phenomenon. It can be carried out with a balance and a freeze-dryer, and has the additional advantage that a large number of candidate formulations can be evaluated simultaneously.

Optimization of storage stability of lyophilized actin using combinations of disaccharides and dextran.

The storage stability of a dry protein depends on the structure of the dried protein, as well as on the storage temperature relative to the glass transition temperature of the dried preparation. Disaccharides are known to preserve the native conformation of a dried protein; however, the resulting T(g) of the sample may be too low ensure adequate storage stability. On the other hand, formulations dried with high molecular weight carbohydrates, such as dextran, have higher glass transition temperatures, but fail to preserve native protein conformation. We tested the hypothesis that optimizing both protein structure and T(g) by freeze-drying actin with mixtures of disaccharides and dextran would result in increased storage stability compared to actin dried with either disaccharide or dextran alone. Protein structure in the dried solid was analyzed immediately after lyophilization and after storage at elevated temperatures with infrared spectroscopy, and after rehydration by infrared and circular dichroism spectroscopy. Structural results were related to the polymerization activity recovered after rehydration. Degradation was noted with storage for formulations containing either sucrose, trehalose, or dextran alone. Slight increases in T(g) observed in trehalose formulations compared to sucrose formulations did not result in appreciable increases in storage stability. Addition of dextran to sucrose or trehalose increased formulation T(g) without affecting the capacity of the sugar to inhibit protein unfolding during lyophilization and resulted in improved storage stability. Also, dextran provides an excellent amorphous bulking agent, which can be lyophilized rapidly with formation of strong, elegant cake structure. These results suggest that the strategy of using a mixture of disaccharide and polymeric carbohydrates can optimize protein storage stability.

Effects of additives on the stability of Humicola lanuginosa lipase during freeze-drying and storage in the dried solid.

The effects of various classes of additives on the stability of a protein with a relatively hydrophobic surface, Humicola lanuginosa lipase (HLL), during lyophilization and storage in the dried solid, were investigated. Prior to lyophilization, it was found that 1 M trehalose or 1% (wt/vol) Tween 20 caused the protein to precipitate. Infrared spectroscopy indicated that trehalose "salted-out" native HLL, whereas Tween 20 induced non-native aggregates. Optimal recovery of native protein in the initial dried solid was obtained in the presence of additives which formed an amorphous phase and which had the capacity to hydrogen bond to the dried protein (e.g., trehalose and sucrose). Additives which crystallized during lyophilization (e.g., mannitol) or which remained amorphous, but were unable to hydrogen bond to the dried protein (e.g., dextran), afforded less stabilization relative to that seen in the absence of additives. Optimal storage stability in the dried solid required that both protein unfolding during lyophilization was minimized and that the formulation was stored at a temperature below its Tg value. Crystallization of sucrose during storage greatly reduced the storage stability of HLL. This was attributed to the increased moisture content and the reduced Tg value in the remaining amorphous phase containing the protein. Sucrose crystallization and the resulting damage to the protein were inhibited by decreasing the mass ratio of sucrose:protein.

Effects of Tween 80 and sucrose on acute short-term stability and long-term storage at -20 degrees C of a recombinant hemoglobin.

The addition of low levels of surfactant polyoxyethylene 20 sorbitan monooleate, Tween 80, to recombinant hemoglobin in phosphate-buffered saline minimized the level of protein aggregation during acute freeze-thaw studies. Addition of sucrose alone to the phosphate-buffered saline formulation, up to 0.5 M, provided minimal protection against freeze-thaw induced aggregation. In contrast to the acute stability studies, long-term storage at -20 degrees C induced aggregation and methemoglobin formation in those formulations containing only Tween 80 in phosphate-buffered saline. Addition of sucrose between 0.1 and 0.5 M to the formulation prevented formation of aggregates and severely arrested methemoglobin formation during the long-term -20 degrees C storage. Specific binding of Tween 80 to the hemoglobin was not observed using 16-doxyl stearic acid partitioning techniques with electron paramagnetic resonance. Minor structural changes to the protein secondary structure during freezing in the absence and presence of Tween 80 were observed with Fourier transform infrared spectroscopy. The alterations were partially prevented by addition of the sucrose. It is likely that the Tween 80 severely reduced protein aggregation during the acute stability studies by preventing the hemoglobin from reaching the air-liquid interface or the liquid-surface interfaces. The reduction in methemoglobin formation and aggregation observed during long-term storage can be accounted for on the premise that the sucrose reduced localized unfolding of the protein in a manner similar to the preferential exclusion theory (Arakawa, T.; and Timasheff, S. N. 1982, Biochemistry 1982, 21, 6536-6544). These studies demonstrate that acute formulation screening studies, albeit useful, may not necessarily predict protein stability during long-term storage.

Effect of zinc binding and precipitation on structures of recombinant human growth hormone and nerve growth factor.

Metal-induced precipitation of protein therapeutics is being used and further developed as a processing step in protein formulation and may have utility in protein purification and bulk storage. In such processes, it is imperative that native protein structure is maintained and the metal complexation is reversible. In the current study, we investigated the effects of zinc-induced precipitation on recombinant human growth hormone (rhGH) and recombinant human nerve growth factor (rhNGF). On the addition of ethylenediaminetetraacetic acid (EDTA), the precipitates were dissolved, yielding complete recovery of native protein in both cases. Both proteins have specific metal binding sites and require specific molar ratios of zinc to protein to initiate precipitation (zinc:rhGH > 2:1; zinc:rhNGF > 18:1). Furthermore, the secondary structures of both proteins were unperturbed in soluble zinc complexes and zinc-induced precipitates, as measured by infrared and circular dichroism spectroscopies. The soluble zinc complex of rhGH had minor tertiary structural alterations, whereas zinc binding did not alter the tertiary structure of rhNGF. These studies indicated that metal-induced precipitation provides a method to maintain proteins in their native state in precipitates, which may be useful for purification, storage, and formulation.

Tween protects recombinant human growth hormone against agitation-induced damage via hydrophobic interactions.

In the absence of surfactants, recombinant human growth hormone (rhGH) rapidly forms insoluble aggregates during agitation. The nonionic surfactant Tween 20, when present at Tween:protein molar ratios >4, effectively inhibits this aggregation. Differential scanning calorimetry (DSC) of rhGH solutions showed melting transitions that decreased by ca. 2 degrees C in the presence of Tween. Circular dichroism (CD) studies of the same thermal transition showed that the decrease is specific to the relatively high protein concentrations required for DSC. CD studies showed melting transitions that decreased with lower protein concentrations. Tween has an insignificant effect on the melting transition of rhGH at lower protein concentrations (0.18 mg/mL). Injection titration microcalorimetry showed that the interaction of Tween with rhGH is characterized by a weak enthalpy of binding. For comparison, interferon-g, another protein which has been shown to bind Tween, also shows weak enthalpy of binding. Fluorescent probe binding studies and infrared spectroscopic investigations of rhGH secondary structure support suggestions in the literature (Bam, N. B.; Cleland, J. L., Randolph, T. W. Molten globule intermediate of recombinant human growth hormone: stabilization with surfactants. Biotechnol. Prog. 1996. 12, 801-809) that Tween binding is driven by hydrophobic interactions, with little perturbation of protein secondary structure.

Effect of Tween 20 on freeze-thawing- and agitation-induced aggregation of recombinant human factor XIII.

Agitation- and freeze-thawing-induced aggregation of recombinant human factor XIII (rFXIII) is due to interfacial adsorption and denaturation at the air-liquid and ice-liquid interfaces. The aggregation pathway proceeds through soluble aggregates to formation of insoluble aggregates regardless of the denaturing stimuli. A nonionic surfactant, polyoxyethylene sorbitan monolaurate (Tween 20), greatly reduces the rate of formation of insoluble aggregates as a function of surfactant concentration, thereby stabilizing native rFXIII. Maximum protection occurs at concentrations close to the critical micelle concentration (cmc), independent of initial protein concentration. To study the mechanistic aspects of the surfactant-induced stabilization, a series of spectroscopic studies were conducted. Electron paramagnetic resonance spectroscopy indicates that binding is not occurring between Tween 20 and either the native state or a folding intermediate state of rFXIII. Further, circular dichroism spectroscopy suggests that Tween 20 does not prevent the secondary structural changes induced upon guanidinium hydrochloride-induced unfolding. Taken together, these results imply that Tween 20 protects rFXIII against freeze-thawing- and agitation-induced aggregation primarily by competing with stress-induced soluble aggregates for interfaces, inhibiting subsequent transition to insoluble aggregates.

Aggregation of recombinant human interferon gamma: kinetics and structural transitions.

Protein aggregation is a complex phenomenon that can occur in vitro and in vivo, usually resulting in the loss of the protein's biological activity. While many aggregation studies focus on a mechanism due to a specific stress, this study focuses on the general nature of aggregation. Recombinant human interferon-gamma (rhIFN-gamma) provides an ideal model for studying protein aggregation, as it has a tendency to aggregate under mild denaturing stresses (low denaturant concentration, temperature below the Tm, and below pH 5). All of the aggregates induced by these stresses have a similar structure (high in intermolecular beta-sheet content and a large loss of alpha-helix) as determined by infrared and circular dichroism spectroscopy. Thermally induced and denaturant-induced aggregation processes follow first-order kinetics under the conditions of this study. Spectroscopic and kinetic data suggest that rhIFN-gamma aggregates through an intermediate form possessing a large amount of residual secondary structure. In contrast to the aggregates formed under denaturing stresses, the salted-out protein has a remarkably nativelike secondary structure.

The effects of Tween 20 and sucrose on the stability of anti-L-selectin during lyophilization and reconstitution.

We have chosen an anti-L-selectin antibody as a model protein to investigate the effects of sucrose and/or Tween 20 on protein stability during lyophilization and reconstitution. Native anti-L-selectin secondary structure is substantially retained during lyophilization in the presence of sucrose (1 or 0.125%). However, aggregation of the protein during reconstitution of lyophilized protein powders prepared without sucrose is not reduced by the presence of sucrose in the reconstitution medium. Aggregate formation upon reconstitution is completely inhibited by freeze drying the protein with sucrose and reconstituting with a 0.1% Tween 20 solution. Tween 20 (0.1%) also partially inhibits loss of native anti-L-selectin secondary structure during lyophilization. However, upon reconstitution the formulations lyophilized with Tween 20 contain the highest levels of aggregates. The presence of Tween in only the reconstitution solution appears to inhibit the transition from dimers to higher order oligomers. Potential mechanism(s) for the Tween 20 effects were investigated. However, no evidence of thermodynamic stabilization of anti-L-selectin conformation (e.g., by Tween 20 binding) could be detected.

Preparation and in vitro characterization of gentamycin-impregnated biodegradable beads suitable for treatment of osteomyelitis.

A new method for preparing poly(L-lactide) (PLA) biodegradable beads impregnated with an ionic aminoglycoside, gentamycin, is described. The process employs hydrophobic ion pairing to solubilize gentamycin in a solvent compatible with PLA, followed by precipitation with a compressed antisolvent (supercritical carbon dioxide). The resulting precipitate is a homogeneous dispersion of the ion-paired drug in PLA microspheres. The microspheres are approximately 1 microm in diameter and can be compressed into beads (3-6 mm in diameter) strung on surgical sutures for implantation. The bead strings exhibit no significant change in release kinetics upon sterilization with a hydrogen peroxide plasma (Ster-Rad). The kinetics of gentamycin release from the PLA beads are consistent with a matrix-controlled diffusion mechanism. While nonbiodegradable poly(methyl methacrylate) (PMMA) beads initially release gentamycin in a similar manner, the drug release from PMMA ceases after 8 or 9 weeks, while the PLA beads continue to release drug for over 4 months. Moreover, only 10% of the gentamycin is released from the PMMA beads, while PLA beads release more than 60% of their load, if serum is present in the release medium. The PLA system displays improved release kinetics relative to PMMA, is biodegradable, is unaltered by gas sterilization, can be used for a range of antibiotics, and can be manipulated without disintegration. These are all desirable properties for an implantable drug delivery system for the prevention or treatment of osteomyelitis.