*2.2.1.2 Powder output*

The powder output (% *w/v*) was defined based on the spray-dried powder (g) collected per input liquid solution (mL). The amount of input liquid solution was 50 mL for each run; the exact volume was recorded and considered for the powder output calculations. The vessel collecting the dry powder from each run was weighted before and after the run, to determine the amount of collected powder from the run. For example, if 4.9 g of powder was generated from 50 mL of input liquid, the powder output was determined as 9.7%.

#### *2.2.1.3 Reconstitution of powder*

For the reconstitution score assessment, the solid powder sample was dissolved in the solvent (Milli-Q® water) to a dry weight percentage of 8% *w/v.* This value resembled the approximate dry weight (% *w/v*) of the input liquid, when comparing the product output across the samples. This was kept consistent between sample runs by dissolving 200 mg powder in 2.5 mL. Upon reconstitution, the solution was left undisturbed for 15 minutes and then vortexed briefly to obtain a homogenous state. The solution was then visually inspected and ranked based on the solubility performance, the intensity of cloudiness, and color formation. The solubility capability of the dissolved powder was ranked from 1 to 4, from bad (1) to good (4) solubility. Low scores (1) were given to cloudy solutions with visible particles, and high scores (4) were given to solutions that were clear and fully homogenous with no visible particles. Solutions from the reconstitution assessment were used further for the quantifications of the growth factor (see Section 2.2.1.4) and the total protein (see Section 2.2.1.5).

#### *2.2.1.4 Growth factor quantification*

For *h*EGF/powder (μg/mg) assessment, the amount of *h*EGF target protein was quantified by capillary-based nano immunoassay, JESS Simple Western™ (*ProteinSimple®, Bio-Techne, Minneapolis, MN, USA*). The *h*EGF amount was quantified within each reconstituted sample based on a generated standard curve of *h*EGF with a known *h*EGF concentration and the concentration of *h*EGF within the powder, which could then be calculated compared to the standard. The *hEGF*/powder (μg/mg) value indicates the product yield and stability of the growth factor after the spray drying step. Receiving a low concentration can, for example, indicate aggregation or irreversible protein denaturation as an indirect estimate, such as the *h*EGF not being able to dissolve after spray drying. Refer to **Appendix A** for a more detailed method description for the growth factor quantification.

#### *2.2.1.5 Total protein quantification*

Total protein/powder (μg/mg) was measured with Bradford assay. Since the growth factor is semi-purified, other barley proteins are also present in the final product, and these need to be considered. Therefore, native barley proteins were quantified, along with the recombinant growth factor. Reduction of the total protein within the powder can, for example, indicate that the protein is forming an insoluble material due to the spray drying process. Refer to **Appendix A** for a more detailed method description for the total protein quantification.

#### *2.2.2 Optimization: Analysis of DoE data*

The data were fit with an RSM model with linear regression (Eq. (1)), and model reduction was performed by enforcing a 95% confidence interval, including only factors and interactions with a p-value <0.05 and factors containing statistically significant effects; refer to **Table 3**. As with the factors, a 95% confidence level was chosen for the responses in the model. The responses in the RSM model that had p-value <0.05 were considered statistically significant; others were disregarded. This RSM model can only predict outcomes for the responses "powder output" and "reconstitution score."

$$Y = a + \sum\_{i=1}^{3} b\_i X\_i + \sum\_{i=1}^{3} c\_i X\_i^2 + \sum\_{i=1}^{2} \sum\_{j=i+1}^{3} d\_{ij} X\_i X\_j \tag{1}$$

*Y* is the yield; *a* is the intercept; *bi, ci*, and *dij* are model coefficients; and *Xi* and *Xj* represent the model regressors.

A single sample, *sample #3*, was eliminated from the DoE analysis due to a loss of powder to the cyclone, for technical reasons. However, since this error did not influence the bioactivity of the sample, *sample #3* was included in the following bioactivity measurements and the SEM analysis. All other samples and data were included, and no outliers were detected. Factors and response values are summarized below in **Table 4**.

The variation in the amount of *h*EGF and total protein per powder, respectively, could not be explained by the factors included in the model. This confirms that the spray dryer settings, even at extremities, do not affect either the growth factor quantity or the overall total protein quantity. This data shows the robustness of the spray drying process within the tested ranges.


#### *Outlet temperature (°C), Spray gas (L/h), and FFR (mL/min). ˆ denotes factors with containing effects above them.*

#### **Table 3.**

*Effect summary. Factors and interactions that influence the responses in the RSM model and their corresponding P values.*


*Excipient-Free Spray Drying of Bioactive Recombinant Proteins Produced in Plants DOI: http://dx.doi.org/10.5772/intechopen.112944*

#### **Table 4.**

*Experimental design matrix, containing input factors and output responses from all sample runs.*

Powder output and reconstitution score were found to be statistically significant, and the model can therefore explain the variation in their responses. The outlet temperature affected reconstitution the most, followed by the interactions of *FFR x outlet temperature* and *FFR x spray gas flow rate*. For the powder output, the interactions of FFR x outlet temperature and *FFR x spray gas flow rate* were the most impactful. **Figure 2a** illustrates factor interactions.

A quadratic effect was observed in the model for both the spray gas flow rate and the outlet temperature. A quadratic effect in a statistical model means that an optimum has been observed in the defined experimental space. This can be visualized by a curvature in their responses, as illustrated in **Figure 2b.** Optimum settings, within the tested range, were found for outlet temperature and spray gas flow rate, as curvature was observed in their responses. The quadratic effect for outlet temperature was only significant for powder output, and the quadratic effect for spray gas was only significant for reconstitution.

The optimized settings of the input factors to maximize the responses in this case study were found to be the following:

> **Feed Flow Rate** ð**FFR**Þ: 4mL*=*min **Spray Gas Flow Rate**: 1400L*=*h **Outlet Temperature**: 60°C

**Figure 2.**

*Effects of the factors on the responses. (a) Interaction profilers for powder and reconstitution show how the different factors interact to affect the responses. (b) Prediction profiler for maximized responses and optimal setting.*

The software calculated the maximized spray gas flow rate settings at 1800 L/h, but we took economic considerations into account when selecting the optimized settings and lowered the selected spray gas flow rate settings to 1400 L/h.

The RSM approach of the DoE in this case study was successful and provided optimum parameter settings for spray drying growth factors in a stabilizing barley matrix with high quality powder and had no impact on the quantity of the target protein or the barley proteins.

#### *2.2.3 Particle morphology with scanning electron microscope (SEM)*

Particle morphology analysis was carried out to investigate whether there was a correlation between microscopical particle shapes and the applied spray drying conditions. The surface morphology of the spray-dried particles was examined using a field emission scanning electron microscope (FE-SEM), Supra 25 by Zeiss (*Oberkochen, Germany*). Powder samples were mounted to a sample stub and gold coated. Samples were scanned at a voltage of 3.0 kV, and their images were captured at two magnification levels, 1000 and 5000.

Selection criteria for the sample runs ultimately taken for SEM characterization were based on analyzing the morphology of the samples expected to have experienced the upper and lower limits for the outlet temperature, reconstitution rating, and particle size.

The analysis of the SEM images served as a qualitative assessment for the given range of parameters in the study. Morphology classification for spray-dried particles as suggested by Prinn et al. [39] can be divided into four categories: (*I*) smooth

*Excipient-Free Spray Drying of Bioactive Recombinant Proteins Produced in Plants DOI: http://dx.doi.org/10.5772/intechopen.112944*

#### **Figure 3.**

*SEM images of the spray-dried particles at magnification 1000. Shown are captures from the following samples: (a) Sample #9, expected to have the smallest particles, here shown as several raisin-like particles crumpled together. (b) Sample #4, expected to have the largest particles, here shown as a mixture of larger, smooth spheres; collapsed particles; and wrinkled particles.*

spheres, (*II*) collapsed or dimpled particles, (*III*) wrinkled or raisin-like particles, and (*IV*) highly crumpled or folded structures.

The study design range demonstrated a distribution of most of the abovementioned different shapes but mostly showed smooth spheres and raisin-like structures. Shown in **Figure 3** are the samples expected to exhibit the smallest particles (sample #9; 4 ml/min, 100°C and 1800 *l/h*) and the largest particles (sample #4; 7 ml/min, 60°C and 600 *l/h*). By comparing these two different particle formations, these reveal only raisin-like particles for the smallest ones (**Figure 3a**)*,* whereas the largest ones show a combination of the two morphologies, smooth spheres and raisin-like particles (**Figure 3b**). The smaller particles seem to tend to clump together. The other tested formulations showed varying amounts of large smooth spheres with smaller particles always forming raisin-like structures with some indications that the higher outlet temperature results in more crumbled structures compared with lower outlet temperatures where smooth surface is more prevalent (data not shown). In other ways, the SEM analysis reveals that particle morphology is not susceptible to changes within the selected range of process parameters.

#### *2.2.4 Bioactivity of growth factor*

A cell proliferation assay for *h*EGF using 3 T3 fibroblast cells was performed by SBH sciences (*Natick, MA)* to measure the biological activity of the growth factor for selected samples. The cells were seeded on multi-well plates and incubated with a dilution series of a commercial growth factor standard. After a pre-defined incubation period, cell proliferation or cell death had been measured using a colorimetric assay. The biological activity of *h*EGF was expressed as *ED50* (effective dose), which is the concentration of the growth factor that induces 50% of the maximum assay response. Thus, the lower the *ED50* value, the higher the activity.

The sample selection for the bioactivity analysis is summarized in **Table 5.** This assay was performed to determine whether there was a correlation between biological activity and the other output responses investigated in the case study.

The bioactivity curves of all samples were tested for parallelism to determine whether the samples were statistically different from each other. Parallelism was


**Table 5.**

*Sample runs selected for the bioactivity measurement and the resulting ED50 values.*

determined by an F-test using JMP Version 17.0 (*SAS Institute Inc., Cary, NC*) software; see **Appendix B**.

All samples were found to retain their bioactivity after spray drying. Sample #2 was found to have the highest ED50 value, and it was found to be different from the other samples, except for sample #3. Sample #2 had the most extreme settings for all the factors. This indicates that the combination of the extreme process parameters might decrease the bioactivity of spray-dried growth factors. This, however, needs further investigation.

In summary, the biological activity of the growth factor was not disrupted for any of the applied process parameters. Spray drying is a robust process to dry recombinant growth factor solutions while preserving biological activity.
