Article Cite This: Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Characterizing the Surface Roughness Length Scales of Lactose Carrier Particles in Dry Powder Inhalers Bernice Mei Jin Tan, Lai Wah Chan, and Paul Wan Sia Heng* GEA-NUS Pharmaceutical Processing Research Laboratory, Department of Pharmacy, National University of Singapore, 18 Science Drive 4, Singapore 117543
ABSTRACT: Surface roughness is well recognized as a critical physical property of particulate systems, particularly in relation to adhesion, friction, and flow. An example is the surface property of carrier particles in carrier-based dry powder inhaler (DPI) formulations. The numerical characterization of roughness remains rather unsatisfactory due to the lack of spatial (or length scale) information about surface features when a common amplitude parameter such as average roughness (Ra) is used. An analysis of the roughness of lactose carrier particles at three different length scales, designed for specificity to the study of interactive mixtures in DPI, was explored in this study. Three Ra parameters were used to represent the microscale, intermediate scale, and macroscale roughness of six types of surface-modified carriers. Coating of micronized lactose fines on coarse carrier particles increased their microroughness from 389 to 639 nm while the macroroughness was not affected. Roller compaction at higher roll forces led to very effective surface roughening, particularly at longer length scales. Changes in Ra parameters corroborated the visual observations of particles under the scanning electron microscope. Roughness at the intermediate scale showed the best correlation with the fine particle fraction (FPF) of DPI formulations. From the range of 250 to 650 nm, every 100 nm increase in the intermediate roughness led to ∼8% increase in the FPF. However, the effect of surface roughness was greatly diminished when fine lactose (median size, 9 μm) of comparable amounts to the micronized drug were added to the formulation. The combination of roughness parameters at various length scales provided much discriminatory surface information, which then revealed the “quality” of roughness necessary for improving DPI performance. KEYWORDS: dry powder inhaler, surface roughness, adhesion, length scale, digital filtration, fine particle fraction
■
INTRODUCTION
lactose monohydrate crystals can generally be described as regular, flat, and smooth. As the same surface is magnified and observed over shorter length scales, it could be described as much rougher due to the appearance of many small surface irregularities. The otherwise innocuous protuberances become highly visible with magnification. Roughness is therefore a scale-dependent property of the surface. A surface topology is a three-dimensional composite profile of surface asperities of different spacings and amplitudes. Surface roughness is often quantified using a single Ra (mean arithmetic roughness) value, an amplitude parameter that is
Surface roughness (or rugosity) is an apparently simple inherent property of lactose carrier particles used in dry powder inhaler (DPI) formulations. For interactive mixtures where fine respirable particles (1−5 μm) are physically blended with coarse carrier particles (40−200 μm), carrier surface roughness is a primary influence of the physical interactions between the drug and carrier. Roughness affects the interparticulate friction when particles are in motion1 during mixing or aerosolization. In respiratory drug delivery, the distribution and adhesion of fine drug particles on the DPI carrier are strongly governed by the surface properties of carrier particles, in particular their surface roughness.2,3 Depending on the scale of observation (or degree of magnification of surface viewed), the “smoothness” or “roughness” of the surface can vary significantly. The gross surface morphology of sieved α© XXXX American Chemical Society
Received: Revised: Accepted: Published: A
January 2, 2018 February 19, 2018 February 28, 2018 February 28, 2018 DOI: 10.1021/acs.molpharmaceut.8b00007 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Article
Molecular Pharmaceutics
tion of these lactose carriers, the relevance of these Ra parameters to DPI performances could be verified experimentally. Finally, the correlations between roughness at different length scales and the FPF of DPI formulations were investigated in order to better define the “quality” of roughness that should be present on the surfaces of lactose carrier particles so that it can be reliably related to DPI performance.
one-dimensional in nature. The Ra is calculated as the arithmetic mean surface height of the roughness profile. Ra values are often not as informative as visual observation of images obtained through optical or scanning electron microscopes (SEM). In order to characterize surface features at different length scales, a combination or hybrid of amplitude and spatial parameters has been proposed.4 Examples of spatial parameters available from modern surface scanning instruments include the texture aspect ratio (Str) and texture direction of surface (Std) while hybrid parameters include the root-meansquare gradient of the surface (Sdq) and developed area ratio (Sdr). Despite the many roughness parameters proposed, Ra is still most preferred because it is intuitively simple, easy to calculate, and widely reported in many other research articles.5 However, one important reason for the limited applicability of Ra is the lack of length scale information accompanying it. It is possible for two surfaces to possess identical Ra values but have drastically different physical appearances and functional properties. Without information on the length scale of roughness, formulation scientists are often unable to realistically visualize or properly quantify the spatial distribution of surface features on carrier particles.6 Surface features within a spectrum of length scales can exist on a carrier particle, ranging from the micrometer-sized surface protrusions to the nanoscale surface asperities. Researchers have acknowledged the importance of roughness across different length scales and thus proposed carrier surface models to explain the impact of roughness on drug adhesion and aerosolization.7−9 The length scales in these surface models are usually in relation to the particle size of drug. Compared to other physical characteristics of lactose carriers such as the purity, crystallinity, particle size, and specific surface area (SSA), surface roughness is not routinely measured or reported as part of carrier quality specifications.10 Literature reviews on the effects of carrier surface roughness on fine particle fraction have shown that there is neither conclusive evidence nor a clear consensus among researchers on the optimal smoothness or roughness of carrier particles.11,12 This may be due to the differences in characterization tools and parameters used for roughness measurement. It is also extremely difficult to analyze the effects of surface roughness independent of other carrier properties such as particle shape, surface crystallinity, and surface chemistry. The correlation between roughness and DPI performance is therefore not sufficiently robust for it to be adopted as a critical raw material attribute in DPI products. Furthermore, a complete characterization of surface roughness should ideally involve the entire surface area of particles. This can be a complex endeavor which is too time-consuming and impractical to implement for routine use. Thus, alternatives or surrogate measurements of surface roughness such as the SSA13,14 are often used instead. The main objective of this study was to explore the analysis of carrier particle roughness at different length scales (microscale, intermediate-scale, and macroscale roughness) which are relevant to the study of drug−carrier interactions in DPI interactive mixtures. Three Ra parameters were proposed together with their accompanying length scale information so that some degree of spatial orientation of surface features can be discerned. These Ra parameters were used to evaluate several types of surface-modified carriers processed by fluid-bed coating and roller compaction. By reconciling the changes in the Ra parameters with the visual information and the “expected” effects after surface modifica-
■
EXPERIMENTAL SECTION Materials. Lactose 100 M (SpheroLac 100), lactose 200 M (GranuLac 200), and an inhalation grade lactose carrier (InhaLac 230) were obtained from Meggle Pharma, Germany. Micronized lactose and fine lactose were produced separately by milling lactose 100 M particles in a jet mill, combined with a powder classifying system (100 AFG, Hosokawa Micron, Germany). Micronized isoniazid (Taizhou Tianrui Pharmaceutical, China) with median size of 2.5 μm was produced by milling in the jet mill and used as the model drug in all DPI formulations. Isopropyl alcohol (IPA) was supplied by Aik Moh, Singapore. Surface Modifications of Lactose Carrier Particles. Surface roughening by fluid-bed coating and roller compaction was carried out as described in a previous study.15 For the preparation of the smoothened fluid-bed coated carrier (denoted by FBS), a saturated lactose solution of IPA/water (3:2, v/v) was sprayed onto the feed powder comprising 2 kg of lactose 100 M (median size, 102 μm) in a fluid-bed system (FlexStream, MP-1, GEA Aeromatic Fielder, U.K.). For the roughened fluid-bed coated carrier (denoted by FBR), the same feed powder was used but the spray liquid was a 15%, w/w dispersion of micronized lactose (median size, 2 μm) in the saturated IPA/water solution. The roller compacted carriers were prepared by compacting a blend of lactose 100 M/lactose 200 M (1:1, w/w) at two roll forces in a roller compactor (Pharmapaktor L200/30P, Hosokawa Bepex, Germany). Milled particles derived from flakes produced using 50 kN and 70 kN roll forces were denoted as RC50 and RC70, respectively. Commercially available inhalation grade lactose (INH), lactose 100 M (control), and all surface modified carriers (FBS, FBR, RC50, and RC70) were fractionated using the air-jet sieve (Hosokawa Micron, USA) to produce particles within the range of 63−125 μm. Acquisition of Particle Surface Images. The topographical images of the carrier particle surfaces were acquired by an optical profiler (Wyko NT1100, USA) using the white light vertical scanning interferometry mode. The optical profiler was placed on an air table maintained with 3 bar of air pressure to isolate the instrument from vibrations arising from the surroundings. A Mirau type interferometer with a 20× objective lens was used with a 2.0× field-of-view lens. The numerical aperture of the objective is 0.4, and the optical resolution of the system is 0.75 μm, according to the instrument specifications. The field-of-view was 109 μm × 144 μm, and a captured image is composed of a 480 × 736 array of pixels. Each raw surface image was acquired and displayed on the Wyko Veeco Vision 3.0 software. Due to the irregular shape of carrier particles, captured raw surface images were not always completely filled by the particle surfaces. Subregions, with a size of 20 μm × 20 μm, of the raw surface images were selected and analyzed as described in the next section. Digital Filtration and Derivation of Ra Parameters. Roughness parameters were derived for surface features at 3 length scales and were quantified by the following R a B
DOI: 10.1021/acs.molpharmaceut.8b00007 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Article
Molecular Pharmaceutics
compared to the use of median values alone. Ra values from each carrier type were first sorted into histogram bins with widths of 150 nm, from the range of 0−3000 nm. The frequency in each bin was calculated as a percentage of the total number of roughness values obtained for each carrier batch (N = 150). A series of coordinates with y-value equal to the percentage frequency and x-value equal to the midpoint of the interval was obtained and fitted to the log-normal distribution:
parameters; Ra,mic, Ra,int, and Ra,mac quantitatively represented the microscale, intermediate-scale, and macroscale roughness of carrier particles, respectively. Ra,all refers to the Ra which would be typically obtained in a roughness measurement where the length scales of surface features are not considered. This latter parameter served as a reference which contained much “embedded” surface information that would otherwise be lost if no further analysis was applied to extract the underlying surface features. The length scales of roughness and the corresponding spatial frequency ranges for obtaining each roughness parameter are shown in Table 1.
2
y = A ·e−(log x − B / c)
where A, B, and c are constants in the equation. The goodnessof-fit R2 of each distribution was calculated. The log-normal distribution was selected over the normal distribution as Ra parameters were found to be better approximated by the lognormal distribution (all R2 values ≥ 0.9; data not shown). Qualitative comparisons between the methods for surface roughening could then be made by analyzing the shifts in Ra distributions and the spread of Ra values. In Vitro Drug Deposition Studies To Determine Fine Particle Fraction. DPI formulations, prepared with the 6 types of coarse carriers, comprised 2%, w/w of micronized drug and various concentrations of fine lactose (0, 1, 3, and 8%, w/w). The median size of fine lactose was 9 μm. The formulations were prepared by weighing out the appropriate quantities of micronized drug, coarse carrier, and fine lactose (where applicable), amounting to a total mass of 1.5 g, into a glass tube. The contents of the tube were blended in a vortex mixer (WhirliMixer, Fisons Scientific, U.K.) for 10 min. Each DPI capsule (size 3, Qualicaps, Japan) was filled with 30 ± 2 mg of the formulation and loaded into the Rotahaler (Glaxo, U.K.). The in vitro drug deposition experiments in the Next Generation Impactor were performed as described in a previous study.15 As the Rotahaler is a low-resistance inhaler, the maximum recommended airflow rate of 100 L/min was used for the duration of 2.4 s in each test. For each DPI formulation, the test was repeated for 6 capsules. The FPF was determined as the mass percentage of the emitted dose which fell within the aerodynamic size range of 1−6 μm.
Table 1. Roughness Parameters That Represent the Different Length Scales of Interest and Their Corresponding Cutoff Spatial Frequency Ranges Used To Extract Surface Information length scale of interest
length scale range (μm)
Ra parameter
spatial frequency (1/μm)
short medium long overall
0.75−5 5−10 10−20 0.75−20
Ra,mic Ra,int Ra,mac Ra,all
1333−200 200−100 100−50 1333−50
Digital filtration was applied on each subregion image in order to obtain an Ra value at each length scale. A full explanation of the mathematical derivations involved is beyond the scope of this paper. The interested reader can refer to references which introduce the process of digital filtration in roughness analysis.16−19 Briefly, Fourier decomposition was applied to each subregion image (using data from the x, y, and z coordinates) to obtain its constituent sine waves of varying spatial frequencies (i.e., length scales) and amplitudes (i.e., surface heights). The spatial frequency, expressed in 1/μm, characterizes the length scales on the surface and is analogous to frequencies on the time scale expressed in 1/s. Surface features of lower spatial frequencies are more widely spaced while those of higher frequencies are more closely spaced. Collectively, these constituent sine waves can again be used to reconstruct the surface profile with minimum loss of information. By applying the appropriate cutoff values of spatial frequencies according to Table 1, the surface profile of the length scale of interest can then be mathematically extracted. The Ra (mean arithmetic roughness) for each extracted profile was derived as follows: m
Ra =
■
RESULTS AND DISCUSSION Differential Surface Information Indicated by Ra Values at Different Length Scales. Ra values calculated from length scales of 0.75−5 μm, 5−10 μm, and 10−20 μm are described as the microroughness (Ra,mic), intermediate-roughness (Ra,int), and macroroughness (Ra,mac) of carrier surfaces, respectively. These length scales of interest were selected based on the (a) size of respirable drug particles (1−5 μm) and (b) length scales at which the probable physical interactions would occur between drug and carrier surfaces. Generally, crevices and protuberances on the carrier surface that are smaller than the diameters of drug particles constitute the microroughness and can be characterized by Ra,mic. These surface asperities establish a limited number of contact points between adhering drug particles and carrier surfaces and potentially determine both the true area of contact and their distance of separation.20 The lower scale limit of 0.75 μm for Ra,mic was set to coincide with the resolution of the optical system used while the upper limit of 5 μm was related to the largest size of respirable particles. It should be noted in this study that surface features at length scales of 0.75−5 μm were considered as the finest surface detail (i.e., microscale) that could be imaged due to the limited spatial resolution of the optical profiler operating under the white light
n
1 ·∑ ∑ |Zij| m·n i = 1 j = 1
(2)
(1)
where m and n are the number of pixels in the image array in the i and j directions, respectively, and Z is the surface height relative to the mean surface. The mean surface is a plane running centrally through the surface that divides the profile equally above and below the plane. Thus, the Ra value gives an indication of the average height deviation of asperities and crevices on the surface. Digital filtration and calculation of all Ra parameters were performed using the Vision 3.0 software. For each Ra parameter, a total of 150 values were obtained for each type of lactose carrier. Fitting of Roughness Distributions. Since the Ra values were obtained from a small population of particles (similar to particle sizing by microscopy), their frequency distributions could be plotted to enable better visualization of the data set C
DOI: 10.1021/acs.molpharmaceut.8b00007 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Article
Molecular Pharmaceutics
Figure 1. Scanning electron microscope images which show the typical particle morphologies of each of the 6 types of lactose carriers.
3 different length scales could not be clearly distinguished from the images. There were also clusters of both irregular and rough particles along with smooth tomahawk-shaped particles in both carriers. The roller compacted carriers were noticeably different from the other 4 types of carriers due to their high macroroughness and granular appearance. While the imaging of particles (even if at different magnifications) provided very useful qualitative information, more precise quantitative differences in surface roughness could only be investigated through the Ra parameters. The Ra,all and Ra values at 3 length scales from each type of carrier are presented in Table 2. The frequency distributions of Ra at various length scales are presented in Figure 2. The
interferometry mode. Other precise instruments such as the atomic force microscope are able to spatially resolve nanometer-scale surface asperities but suffer from very slow scan speeds, highly limited scan areas, and an inability to image very rough surfaces. The noncontact nature of measurement, high scan speeds, and large area of scans by the optical profiler were more important considerations for roughness measurements in this study. The intermediate scale and macroscale roughness comprise surface crevices wide enough to entrap drug particles/ agglomerates.21 The upper scale limit of 20 μm for Ra,mac was the same as the maximum dimension of the subregion images used in Ra calculations. The qualitative differences in surface roughness can be visually observed in the SEM images of the 6 types of carriers in Figure 1. There were clear differences in roughness across the different length scales. Even after surface roughening, flat and smooth crystal faces were preserved in the FBR carrier particles. The same morphological features were also found in the INH, control, and FBS carriers. Hence, these particles should have relatively low Ra,mac. The surfaces of the FBS carrier particles were visibly smoothened and the edges of the crystals appeared rounder compared to the INH or control carriers, suggesting that the Ra,mic would be low. No reliable comparisons could be made between the control and FBR carriers as roughness at the
Table 2. Surface Roughness of Carrier Particles at 3 Different Length Scales median Ra (nm)
D
Ra param
INH
control
FBS
FBR
RC50
RC70
Ra,mic Ra,int Ra,mac Ra,all
312 231 221 416
389 298 352 617
348 232 221 493
639 356 339 927
670 452 427 1060
1025 640 697 1235
DOI: 10.1021/acs.molpharmaceut.8b00007 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Article
Molecular Pharmaceutics
higher Ra,all of the FBS carrier could be attributed to the wider distribution of Ra,mic, which was also reflected in its higher median Ra,mic value. The control and FBR carriers possess similar roughness at the intermediate scale and macroscale, as suggested by similarities in their Ra,int and Ra,mac. However, the FBR carrier had significantly higher Ra,mic, which was probably due to adhering micronized lactose particles. This resulted in approximately 300 nm increase in the Ra,all. On the other hand, the FBR and RC50 carriers have similar Ra,mic while the RC50 carrier had greater Ra,int and Ra,mac, resulting in the slightly increased Ra,all. Finally, the RC70 carrier showed significantly higher roughness at all length scales, especially at the microscale, giving rise to the highest Ra,all value. Thus, differentiation of Ra parameters according to length scales of interest enabled better rationalization of differences in Ra,all and allowed the spatial distribution of roughness to be more clearly discerned. Comparative Differences in Surface Modifications by Fluid-Bed Coating and Roller Compaction. The surface smoothening effect by fluid-bed coating was evaluated by comparing the control to the FBS carrier. Smoothening of the particle surfaces was achieved at all length scales and was particularly effective at the macroscale since the Ra,mac decreased from 352 to 221 nm and its distribution also became visibly narrower. As the length scale decreased further, the surface smoothening effect became less pronounced. Ra,mic was decreased only from 389 to 348 nm after the surface smoothening effort. This suggested that spraying of lactosesaturated water:IPA mixtures was not very effective at bringing about the dissolution of surface asperities/fines that was essential for smoothening of microscale features. The effect of surface roughening by fluid-bed coating was evaluated by comparing the shift of roughness distribution in the control and FBR carriers. In contrast to the surface smoothening process, surface roughening was much more effective over shorter length scales. The increase in roughness was the greatest for Ra,mic (389 to 639 nm) followed by Ra,int (298 to 356 nm) while Ra,mac (352 to 339 nm) was relatively unchanged. It can be seen that roughness due to physical attachment of micronized lactose fines,22 with a known geometric size of about 2 μm, was well characterized by the Ra,mic. There were only small changes to carrier surfaces over scales of 5−20 μm. When considered over shorter length scales, the FBR and RC50 carriers had similar Ra,mic of about 650 nm. This suggested that the physical attachment of micronized lactose on coarse lactose crystals resulted in comparable microroughness to the RC50 carrier. However, the main advantage of surface roughening by roller compaction over fluid-bed coating was in its creation of surface roughness at the longer length scales. This was observed from the significant increase in Ra,int and Ra,mac of the roller compacted carriers from 200−400 nm to 400−700 nm. Particle fragmentation and bonding were required to create more irregular and granular carrier particles with rough surfaces over the scales of 5−20 μm. At the same time, the increase of the microroughness was very significant in the RC70 carrier compared to the RC50 carrier. The Ra,mic of the RC70 carrier increased from 670 to 1025 nm while the Ra,int and Ra,mac increased by 200−250 nm compared to the RC50 carrier. This suggests that the roll compaction force of 70 kN resulted in extensive fragmentation of lactose particles to generate fine particles which were again bonded at the particle surfaces. While the magnitude of median Ra suggests the overall degree of carrier roughness, the frequency distribution of Ra
Figure 2. Distribution of Ra parameters of each type of lactose carrier at the length scales of (A) 0.75−5 μm, (B) 5−10 μm, and (C) 10−20 μm.
median Ra values of the INH carrier at all length scales were the lowest among all carriers, suggesting that these particles had the smoothest surfaces. Inhalable grades of crystalline α-lactose monohydrate typically undergo multiple crystallization steps to achieve a high degree of chemical and form purity as well as regularity in particle morphologies and surface characteristics. These crystals possess the classical tomahawk shape associated with monoclinic crystals which possess flat and distinct surfaces, giving rise to very low Ra,int and Ra,mac. In addition, these lactose particles undergo long durations of air-jet sieving into narrow size fractions and their surfaces are generally free from adherents (fine lactose particles) which would contribute to the Ra,mic. The FBS carrier showed almost identical median Ra,int and Ra,mac values and distributions to the INH carrier. The E
DOI: 10.1021/acs.molpharmaceut.8b00007 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Article
Molecular Pharmaceutics
contributed positively to FPF. It is shown that the FPF was improved from about 25% when the smoothest carriers were used to about 55% for the roughest carriers. There were relatively good correlations for all Ra parameters (R2 ≥ 0.8) with FPF, but the best correlation was observed for Ra,int. This correlation R2 was greater than that of the Ra,all, suggesting the benefit of analyzing particle roughness at different length scales. The steeper gradients of the best-fit lines for Ra,int and Ra,mac suggest that further increase in the intermediate scale and macroscale roughness is more effective in improving the FPF than the microroughness. There are two observations which support this finding. The control and FBR carriers have similar intermediate scale and macroscale roughness while significantly different microscale roughness. The FPF values of formulations containing the two carriers are similar. The RC50 and FBR carriers have similar microscale roughness but different intermediate scale and macroscale roughness. DPI formulations containing the RC50 carrier have significantly higher FPF. Qualitative differences in roughness can be more clearly observed in the magnified view of particle surfaces of the INH and RC70 carriers in Figure 4. These two carriers had the lowest and highest Ra values at all length scales. The high roughness of the RC70 carrier at the intermediate scale and macroscale shows many locations at which drug particles can potentially adhere, either individually or in the form of smaller agglomerates. However, drug particles/agglomerates tend to accumulate at the sparsely located crevices on the smooth INH particles. These crevices (or areas of roughness) comprise a much smaller surface area compared to the total available surface area of the smooth particles. One major function of the coarse lactose carrier is to break up cohesive agglomerates of micronized drug so that individual drug particles (or smaller drug agglomerates) can be distributed on the carrier surfaces in the interactive mixture. This will help to lower the energies required to detach drug particles from the carrier surface and disperse drug agglomerates during inhalation. Roughness on carrier surfaces may promote this function either by physically increasing the available surface area for drug adhesion or through increased frictional effects during blending. Given that Ra,int showed the best correlation with the FPF of DPI formulations, a graph of FPF of DPI formulations containing fine lactose was plotted against Ra,int in Figure 5. One best-fit line was constructed, using the method of leastsquares, for each set of data corresponding to the particular fine lactose content used (i.e., 0, 1, 3, and 8%, w/w). The gradient of the best-fit line, denoted by m, represented the change in the
shows the variability of surface roughness among carrier particles of the same batch. Similar to particle size, it is necessary to analyze surface roughness with reference to its distribution of values as it is the total particle surface area that affects blending and drug aerosolization behavior. The INH and FBS carriers comparatively showed the narrowest distributions in all Ra parameters among all carriers. Once the lactose particles were subjected to a surface roughening process (i.e., FBR, RC50, and RC70 carriers), the Ra distributions were significantly broadened at all length scales. This showed that the surface roughening process had relatively nonuniform effects at the scale of individual particles. This was most clearly seen in the RC70 carrier where >25% of Ra,mic values were greater than 1500 nm, suggesting the creation of very rough carrier particles even though the median Ra,mic was 1025 nm. When taken together, the shifts in Ra values and widening of Ra distributions appeared to be in good agreement with the nature of the surface roughening processes. For example, the adhesion of micronized lactose particles led to significant increases in the microroughness while roller compaction was required to further increase the intermediate roughness and macroroughness. The changes in Ra at different length scales corroborated the visual information provided by SEM images. Relation of Ra Parameters to the Fine Particle Fraction of DPI Formulations. Figure 3 shows the variation of the FPF
Figure 3. Relationship between different Ra parameters of lactose carriers with the fine particle fraction of DPI formulations.
of DPI formulations with carrier roughness at different length scales. The correlation between each Ra parameter and the FPF is indicated by the R2 value. The positive gradients of all best-fit lines clearly showed that increased surface roughness
Figure 4. SEM images showing the magnified view of carrier particle surfaces in the INH and RC70 carriers, with 10 μm scale lines indicated. The images were taken after the coarse carriers were blended with 2%, w/w of micronized drug. F
DOI: 10.1021/acs.molpharmaceut.8b00007 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Article
Molecular Pharmaceutics
good agreement with visual observations and the theoretical expectations of the surface modification methods employed. Among the various roughness parameters, Ra,int gave the best correlation with the FPF of DPI formulations. However, the effect of carrier surface roughness diminished greatly when fine lactose of comparable amounts to the micronized drug was added to the formulation. In view of the usefulness of these Ra parameters, researchers are encouraged to give sufficient consideration to the spatial distribution of surface features while attempting to quantify particle surface roughness using the conventional Ra parameters. While such techniques involve many more mathematical manipulations and are tedious to perform, they enable much greater utilization of the threedimensional information recorded by modern surface imaging instruments.
Figure 5. Relationships between median Ra,int and fine particle fraction of DPI formulations containing 0−8%, w/w of fine lactose. The m values are the gradients of the best-fit lines.
■
AUTHOR INFORMATION
Corresponding Author
*Tel: +65 65162930. Fax: +65 67752265. E-mail: phapaulh@ nus.edu.sg.
FPF for every 1 nm increase in the Ra,int of the lactose carrier. The improvement of FPF by carrier roughness was clearly dependent on the fine lactose content. In the absence of fine lactose, every 100 nm increase in the Ra,int resulted in an average of 8% increase in the FPF. A significant decrease in the gradient of the best-fit line, from 0.08 to 0.05, was observed when just 1%, w/w of fine lactose was added. As the content of fine lactose increased to 3 or 8%, w/w, roughness no longer had a significant positive influence on the FPF (m = 0.01 and m = −0.01). This result suggests that the effect of carrier surface roughness on FPF may be greatly diminished when comparable amounts of fine lactose and drug are present. This is because the FPF would be dependent on carrier roughness only when a large proportion of drug particles is directly adhered onto the carrier surfaces. When the relative availability of fine lactose to drug is increased (such as in the presence of 3 or 8%, w/w of fine lactose), the association between drug particles and carrier surfaces would be reduced. On a similar note, Ra,int had a good predictive effect on FPF only when the fine lactose content was between 0 and 1%, w/w. This observation possibly suggests that the extensive association of fine drug particles with fine lactose in mixed drug−lactose agglomerates renders the carrier surface roughness a much less critical influence of drug aerosolization. When the fine lactose content was increased from 3 to 8% (Figure 5), the FPF of all formulations containing the different carriers was increased only by 5 to 7% (except the roughest carrier RC70 where no change was observed). In actual DPI mixtures, however, only a limited amount of fine lactose can be added. This is because cohesive fine lactose has detrimental effects on blend uniformity and powder flow which affects the manufacturability of DPI formulations.
ORCID
Paul Wan Sia Heng: 0000-0002-8354-5764 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors would like to acknowledge the financial support from GEA-NUS PPRL fund (N-148-000-008-001). B.T.M.J. is a recipient of the National University of Singapore President’s Graduate Fellowship.
■
REFERENCES
(1) Dickhoff, B. H.; de Boer, A. H.; Lambregts, D.; Frijlink, H. W. The interaction between carrier rugosity and carrier payload, and its effect on drug particle redispersion from adhesive mixtures during inhalation. Eur. J. Pharm. Biopharm. 2005, 59 (1), 197−205. (2) Thalberg, K.; Berg, E.; Fransson, M. Modeling dispersion of dry powders for inhalation. The concepts of total fines, cohesive energy and interaction parameters. Int. J. Pharm. 2012, 427 (2), 224−233. (3) El-Sabawi, D.; Edge, S.; Price, R.; Young, P. M. Continued Investigation Into the Influence of Loaded Dose on the Performance of Dry Powder Inhalers: Surface Smoothing Effects. Drug Dev. Ind. Pharm. 2006, 32 (10), 1135−1138. (4) ISO, ISO 25178-2:2012. In Geometrical product specifications (GPS) Surface texture: ArealPart 2: Terms, definitions and surface texture parameters, ISO/TC 213: Switzerland, 2012. (5) Karande, A. D.; Zhou, Q.; Chan, L. W.; Heng, P. W. S.; Liew, C. V. Scanning Probe Microscopy for Surface Roughness Characterization of Lactose Carriers Used in Dry Powder Inhalation Therapy. J. Scanning Probe Microsc. 2007, 2 (1−1), 5−9. (6) Tan, B. M. J.; Liew, C. V.; Chan, L. W.; Heng, P. W. S. Particle Surface RoughnessIts Characterisation and Impact on Dry Powder Inhaler Performance. In Pulmonary Drug Delivery; Nokhodchi, A., Martin, G. P., Eds.; John Wiley & Sons, Ltd.: United Kingdom, 2015; pp 199−222. (7) Kawashima, Y.; Serigano, T.; Hino, T.; Yamamoto, H.; Takeuchi, H. Effect of surface morphology of carrier lactose on dry powder inhalation property of pranlukast hydrate. Int. J. Pharm. 1998, 172 (1), 179−188. (8) Heng, P. W. S.; Chan, L. W.; Lim, L. T. Quantification of the surface morphologies of lactose carriers and their effect on the in vitro deposition of salbutamol sulphate. Chem. Pharm. Bull. 2000, 48 (3), 393−8. (9) Littringer, E. M.; Mescher, A.; Schroettner, H.; Achelis, L.; Walzel, P.; Urbanetz, N. A. Spray dried mannitol carrier particles with
■
CONCLUSION An analysis of the surface roughness of lactose carrier particles at 3 different length scales, designed for specificity to the study of interactive mixtures in DPI formulations, was explored in this work. In contrast to most of the published research on carrier surface roughness, all Ra parameters reported here were accompanied by their length scale information. While roughness profiles of interest could be mathematically extracted based on the input of the physical length scales, the relevance of these manipulations could only be established by associating them with visual information from the SEM images and with reference to the in vitro drug deposition results. The changes in the median values and distribution of Ra parameters were in G
DOI: 10.1021/acs.molpharmaceut.8b00007 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Article
Molecular Pharmaceutics tailored surface properties–the influence of carrier surface roughness and shape. Eur. J. Pharm. Biopharm. 2012, 82 (1), 194−204. (10) Center for Drug Evaluation and Research. Metered dose inhaler (MDI) and dry powder inhaler (DPI) drug products: Chemistry, manufacturing and controls documentation; U.S. Department of Health and Human Services: USA, 1998. (11) Lin, Y.-W.; Wong, J.; Qu, L.; Chan, H.-K.; Zhou, Q. Powder production and particle engineering for dry powder inhaler formulations. Curr. Pharm. Des. 2015, 21 (27), 3902−3916. (12) Zhou, Q.; Morton, D. A. V. Drug−lactose binding aspects in adhesive mixtures: Controlling performance in dry powder inhaler formulations by altering lactose carrier surfaces. Adv. Drug Delivery Rev. 2012, 64 (3), 275−284. (13) Hino, T.; Serigano, T.; Yamamoto, H.; Takeuchi, H.; Niwa, T.; Kawashima, Y. Particle design of Wogon extract dry powder for inhalation aerosols with granulation method. Int. J. Pharm. 1998, 168 (1), 59−68. (14) Ferrari, F.; Cocconi, D.; Bettini, R.; Giordano, F.; Santi, P.; Tobyn, M.; Price, R.; Young, P.; Caramella, C.; Colombo, P. The surface roughness of lactose particles can be modulated by wetsmoothing using a high-shear mixer. AAPS PharmSciTech 2004, 5 (4), 69−74. (15) Tan, B. M. J.; Chan, L. W.; Heng, P. W. S. Improving Dry Powder Inhaler Performance by Surface Roughening of Lactose Carrier Particles. Pharm. Res. 2016, 33 (8), 1923−1935. (16) Seah, M.; Chiffre, L. D. Surface and Interface Characterization. In Springer Handbook of Metrology and Testing, 2nd ed.; Czichos, H., Saito, T., Smith, L., Eds.; Springer-Verlag Berlin Heidelberg: Berlin, 2011; pp 281−335. (17) D Antonio, P.; Lasalvia, M.; Perna, G.; Capozzi, V. Scaleindependent roughness value of cell membranes studied by means of AFM technique. Biochim. Biophys. Acta, Biomembr. 2012, 1818 (12), 3141−3148. (18) Blateyron, F. Filtation techniques for surface texture. http://www. digitalsurf.com/en/guide-filtration-techniques.html (26 Aug 2017). (19) Novak, M.; Bui, S. ISO-Standardized filtering for DektakXT Stylus Profilers. Bruker Nano Surfaces Division: USA, 2012. (20) Whitehouse, D. J.; Archard, J. F. The Properties of Random Surfaces of Significance in their Contact. Proc. R. Soc. London, Ser. A 1970, 316 (1524), 97−121. (21) Boshhiha, A. M.; Urbanetz, N. A. Influence of carrier surface fines on dry powder inhalation formulations. Drug Dev. Ind. Pharm. 2009, 35 (8), 904−16. (22) Chan, L. W.; Lim, L. T.; Heng, P. W. S. Immobilization of fine particles on lactose carrier by precision coating and its effect on the performance of dry powder formulations. J. Pharm. Sci. 2003, 92 (5), 975−984.
H
DOI: 10.1021/acs.molpharmaceut.8b00007 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
