2.5.1. Study of PVA-Based Carrier Erosion

The erosion and dissolution of empty PVA-based carriers were characterized by gravimetry, and to trace the number and particle size of PVA colloidal-sized aggregates dispersed in various dissolution media, the DLS (dynamic light scattering) method was utilized.

To evaluate the effect of ionic strength and surface-active agents on degradability, the erosion studies of the empty carriers were conducted in aqueous media of pure demineralized water, pH = 1.2 HCl, pH = 6.8 phosphate buffer and pH = 6.8 tris-(hydroxymethyl) aminomethane (TRIS) buffer with or without sodium salt of bile acids (cholic acid and deoxycholic acid sodium salt 1:1 mixture (Sigma-Aldrich, St. Louis, MO, USA). The erosion of the carriers was tracked visually (Olympus Stylus TG-4 digital camera, Olympus Corp., Tokyo, Japan).

The erosion of the PVA-based carriers was tracked visually.

Filtered erosion samples were measured with the instrument Zetasizer Nano ZS™ (Malvern Instruments Ltd., Malvern, UK) for the derived count rate (DCR) and particle size. The Zetasizer® instrument was equipped with a He-Ne laser (wavelength 633 nm, 4.0 mW) and an avalanche photodiode served as a detector at a detection angle of 173◦ (backscatter mode). Transmittance values for DLS were quantified by Agilent 8453 UV-Visible Spectrophotometer (Agilent Technologies Ltd., Santa Clara, CA, USA) at wavelength 633 nm. Measurement settings: automatic mode, NIBS (none-invasive-back-scattering) 173◦, 30 sub runs/measurements; run duration: 10 secs, automatic laser position selected at 4.65 mm from the bottom of the cuvette; attenuation: attenuator 9 was selected automatically. Three measurements with 30 runs were performed for each sample, and the mean ± SD values are reported for all DLS parameters in this article.

The weight of the 3D printed PVA carriers was determined on an analytical balance (Sartorius LA 230S, Sartorius AG, Göttingen, Germany), which will give the initial weight (*Wi*) during the calculation. The carrier was then placed in the apparatus used in the in vitro dissolution test described above. The test was also performed under the conditions mentioned above. The dissolution medium was pH = 1.2. The printlets were taken out 5, 15, 30, 60, 120, and 240 min later; the dissolution medium was removed by vacuum filtration using a PyrexTM borosilicate glass filter. The residue was stored in an oven (6030 Heraeus Instruments GmbH, Hanau, Germany) at 70◦ C for 48 h. Dry printlet mass (*Wdry*) was then determined (Sartorius LA 230S, Sartorius AG, Göttingen, Germany). The weight loss by erosion of carriers was calculated by Equation (1) respectively [46]:

$$weight\ \left(\%\right) = \frac{w\_{dry}}{w\_i} \times 100\tag{1}$$

*Wdry*—mass of the dried printlet; *Wi*—initial mass of the printlet.

#### 2.5.2. Riboflavin Release

The riboflavin concentrations of the dissolution samples were measured by UVspectroscopy (Agilent 8453 UV-Vis spectrophotometer; Agilent Technologies, Waldbronn, Germany) at 267 nm.

Numerous theories and kinetic models describe and applied for the characterization of drug dissolution profile [47]. Since the shape of the investigated curves was different, the Weibull distribution function (1) was used for the characterization of the dissolution profile of the riboflavin loaded PVA or PLA-based 3D printlet [48].

$$M\_{\rm f} = M\_{\infty} \left[ 1 - e^{-\left(\frac{t - t\_0}{\tau\_d}\right)^{\beta}} \right] \tag{2}$$

*Mt*—the percentage of the dissolved API at time; *M*∞—the infinite concentration of the API in percentages; *t0*—dissolution lag time; *β*—curve shape parameter; *τd*—time in minutes when 63.2% of the API has been dissolved.

Where *Mt* is the percentage of the dissolved active pharmaceutical ingredient at time t, *M*<sup>∞</sup> is the infinite concentration (%) of the drug, *t0* is the dissolution lag-time, *β* is the shape parameter of the curve, and the *τ<sup>d</sup>* represents the time (minutes) when 63.2% of the drug has been dissolved.

#### **3. Results and Discussion**

There are several formulations printed with polyvinyl alcohol, due to its soluble character [21,23,29,32]. However, there is extended research available in connection with the API–PVA formulations [49]; the erosion of PVA itself has not yet been described.

#### *3.1. The CAD Design and the Tracking of the Printlet*

The printing process (Figure 1) was captured by a FLIR thermal camera in order to obtain information about the already printed layers, while the upper ones are printed on the structure. The pinpoint set onto the carrier indicates that while significant heat comes from the nozzle and the heated bed towards the printlet, the state of the solidified layers is extremely acceptable. The temperature of the printed wall has not reached the 50 ◦C, and this phenomenon means that the filling of thermolabile API can be accomplished during the process.

#### *3.2. Physical Characterisation of 3D Printed Carrier Systems*

To guarantee the reproducibility of the manufacturing quality, the investigation of physical characteristics was executed on the printlets. As there is currently no official description in the pharmacopeia for the study of 3D printed carriers, we performed the study according to the pharmacopoeial description (Ph. Eur. 9.) of uncoated tablets.

The measurement results are shown in Table 1. For uniformity of mass, the standard deviation was minimal, well below the 5% allowed in the pharmacopeia. To check the print settings, it is also important to check the height and diameter of the 3D printed carrier. The measured values were close to the original value of the set parameters. The pharmacopoeia concedes a 1% weight loss for friability when testing uncoated tablets. Compared to this value, both PVA and PLA-based printlets had very low friability values, as shown in the data in Table 1. Besides replicability, these results indicate a good opportunity to produce fillable carriers not just for immediate usage but stock can be also piled from them. Due to the structure of the carriers, the weak point of this CAD design is the last layer of the wall around the hollow and the first layer of the closing top area of the printlet. The joint section of the two different layers ruptures if the hardness tests are executed. However, the lowest value of the hardness test performed was 205 N, while the highest was 350 N for PVA. The standard deviation surpasses the 5% limit; however, with the values oscillating in this territory, the mechanical behavior of the carriers shows no diverse differences.

**Table 1.** Physical characteristics of 3D printed carriers.


Digital microscopic images show the one-orifice PVA and PLA-based carrier (Figure 1). It is clear from the images that the layers formed according to the design file during FDM printing processes. The top view shows the designed orifice that plays an important part in the filling and dissolution of the active ingredient. Figure 2 shows the cross-section view of the biodegradable polymer carriers containing multiple drug delivery orifices design for 3D printing and the prepared prototypes.

### *3.3. Erosion of the PVA Carrier*

As earlier mentioned PVA is commonly referred to as a water-soluble excipient, but PVA forms a physical hydrogel in an aqueous medium [50]. Of course, concentration conditions must be taken into account. Dilution of a physical hydrogel with water gives a colloidal solution. During the studies, our aim was to follow the behavior of the printed PVA carrier through the simulated circumstances of the GI tract. The third figure (Figure 3) shows what happens to a PVA-based carrier contacting aqueous medium. The colloidal dissolution/erosion of PVA is a consecutive process. The digital microscope image in the figure shows the wall of a PVA carrier located in a 90-minute release medium (pH = 1.2). Erosion and gel state of PVA can be observed. The solid-state wall is eroded into physical hydrogel state and very small fibers by the medium, forming a gel state before forming the colloidal solution. In the case of water-soluble PVA, the weight of the carrier decreases continuously (black line), while the DCR value determined from the release medium

increases (red line, violet line). It is known in the literature that the increase in DCR is due to an increase in the concentration of the dispersed particles and/or an increase in the size of the dispersed particle [51].

**Figure 3.** Erosion of PVA-based carrier (**A**): black line—weight loss; red line—pH = 1.2 HCl; (**B**): red line—pH = 1.2 HCl; violet line:—pH = 6.8 phosphate buffer (carrier = PVA; n = 3; mean ± SD). (**C**,**D**): digital images of PVA wall during dissolution (pH = 1.2; time = 90 min).

Since the particle size in the samples was below 300 nm (Figure 4) for the entire period, the decisive process was the increase in the concentration of PVA in the dissolution medium. The colloidal dissolved particles reach the colloid particle size interval and remain in this state during at least 24 h.

**Figure 4.** Visual tracking of the PVA-based carrier erosion in different media.

Comparing the derived count rates of each carrier dissolved in different media (Table 2) shows that the highest proportion of the PVA walls are being dissolved during the first 120 min in the case of demineralized water, pH = 1.2 solution. Comparing the derived count rates of each carrier dissolved in different media (Table 2) shows that the highest proportion of the PVA walls are being dissolved during the first 120 min in the

case of demineralized water, pH = 1.2 solution. However, in the TRIS buffer (pH = 6.8), used in order to avoid the PVA incompatibilities with inorganic phosphate [52], the erosion and formation of colloidal solution process were much slower, and the presence of surface active bile salts did not accelerate the progress. The visible tracking of the solid samples also showed that the intact structure of the solid carriers did not disappeare during the first two hours of the study, but as can be seen in the images and in the weight measurement, at 240 min the carrier is completely disintegrated into colloidal particles (Figures 3–5).

**Table 2.** The percentages of derived count rates *(DCRt /DCR1440min* × *100)* of various dissolution media (carrier = PVA; n = 3; mean ± SD).


**Figure 5.** Particle size of dissolution sample (carrier = PVA; n = 3; mean ± SD).
