Influence of Acid and Pepsin on ¹³C-Formaldehyde Induced Gelatin Crosslinks
Influence of Acid and Pepsin on ¹³C-Formaldehyde Induced Gelatin Crosslinks Using Carbon-13 NMR
Thomas B. Gold, Ph.D. and George A. Digenis, Ph.D.
To observe, using carbon-13 nuclear magnetic resonance (¹³C-NMR), the effect of dilute acid and pepsin on gelatin crosslinks, induced by addition of ¹³C-enriched formaldehyde (¹³CH2O) to an aqueous gelatin solution.
Gelatin has been used extensively in the pharmaceutical capsule industry due to its favorable properties: solubility in aqueous solutions, glossy appearance, ease of swallowing, ability to hold dyes and opacifiers, and strong flexible backbone (1). The versatility of gelatin capsules has long been known, as proven by the wide range of formulations which have been incorporated into both soft and hard gelatin capsules (2).
Gelatin is susceptible to chemical modification by a variety of reagents, of which formaldehyde has been examined in greatest detail (3). Crosslinking of proteins by aldehydes has been used extensively by the biological (4,5) and medical sciences (6,7) as well the leather (8) and photographic industries (9). The utility of crosslinking is limited however, in the pharmaceutical industry. Specifically, gelatin capsules which are intended for immediate release and biovailability of their contents are rendered insoluble upon exposure to even trace levels of such aldehydes (10). It has been reported, for example, that corn starch, a common drug excipient and fill material in hard gelatin capsules (HGCs), may contain low levels of hexamethylenetetramine stabilizer (11-14), which decomposes under humid conditions to give ammonia and formaldehyde. Polyethylene glycol ethers, frequently used as formulation excipients for drugs in soft elastic gelatin capsules (SEGs), liberate low molecular weight aldehydes through free radical reactions, upon exposure to aerobic conditions (11,15). Excipient contamination/degradation may contribute to the partial or even total insolubilization of the gelatin shell, preventing the rapid release of drug from the capsule. Because of dissolution problems in in vitro testing of hard and soft gelatin capsules, the potential effects of formaldehyde crosslinking on the in vivo performance of the same capsules have been a matter of concern in situations where gelatin capsules are utilized as delivery systems (1).
Formaldehyde appears to chemically react with both the ε-amino and guanidino funtionalities of lysine and arginine, respectively, on the gelatin polypeptide to form methylols (1). Subsequent crosslinking can occur when the methylols react with either guanidino or amino functionalities to form methylene bridges (1).
Carbon-13 nuclear magnetic resonance spectroscopy (¹³C-NMR) has been utilized to study the formation of reaction intermediates and crosslinked products resulting from the addition of formaldehyde to gelatin solutions (16-19). ¹³C-NMR studies suggested that pancreatin could depolymerize crosslinked gelatin without affecting the methylene crosslink units bridging lysine and arginine within the gelatin polypeptide backbone (19).
Here, we have employed ¹³C-NMR spectroscopy to investigate the proclivity of formaldehyde-crosslinked gelatin toward hydrolysis by pepsin and/or hydrogen ions. The interaction of these two species with crosslinked gelatin, observed in vitro using ¹³CH2O and ¹³C-NMR, may assist in predicting the fate of crosslinked gelatin capsules in gastrointestinal milieu.
- Type A (acid hydrolyzed pigskin) and Type B (lime hydrolyzed bone) gelatin blend used for the manufacture of two piece hard gelatin capsules was obtained from Capsugel, Colmar, France.
- Carbon-13 labeled formaldehyde, ¹³CH2O, 20% aqueous solution, 99+ atom % ¹³C was obtained from Isotec, Inc., Miamisburg, OH.
- Deuterium oxide (99.9%) was obtained from Cambridge Isotope Laboratories, Woburn, MA.
Standard ¹³C-NMR spectra were measured on a Varian VXR-300 (Varian, Palo Alto, CA), using an 8.0 μsec pulse width and a 2.0 acquisition time with 1800 transients.
A solution of 6% gelatin was prepared by allowing 600 mg gelatin powder to swell in 5 mL H2O for 10 min. An additional 3.4 mL H2O and 1.0 mL D2O were added, and the gelatin was dissolved at 38°C. After adjusting the pH of the solution to 7.20 with dilute NaOH, 1.0 mL was transferred to each of three test tubes and two 5 mm NMR tubes. Crosslinking of the five gelatin solutions was induced by addition of 10 μL of 20% ¹³CH2O to each sample, giving 2000 ppm formaldehyde in each solution. Both test tubes and NMR tubes remained capped at ambient conditions for 24 h, after which a ¹³C-NMR spectrum was obtained for the solution within one of the NMR tubes. The four remaining samples were treated with pepsin and/or acid according to the experimental design in Table 1, after which they were capped and stored at 37°C for 24 h. Carbon-13 NMR spectra were obtained for Solution 2 and Solutions 3-5, after transferring the latter to NMR tubes. Seven solutions of 6% gelatin prepared as above were crosslinked by the addition of 10 µl of 20% ¹³CH2O solution to each NMR tube of 1.00 ml of the aqueous gelatin solution. The NMR tubes were left at ambient conditions for 24 h, after which they were treated with 3M HCl (Table 2), capped, and incubated at 37°C for additional 24 h. Standard ¹³C-NMR spectra were then obtained for each of the seven formaldehyde-crosslinked solutions.
Table 1. Experimental protocol for addition of acid and/or pepsin to aqueous solutions of 6% gelatin, crosslinked for 24 h with 2000 ppm ¹³CH2O.
Table 2. Experimental protocol for addition of acid to aqueous solutions of 6% gelatin, crosslinked for 24 h with 2000 ppm ¹³CH2O.
Chemical shift assignments of the ¹³C-formaldehyde-crosslinked gelatin solutions were made according to previously published work (16-19) and were confirmed using semiempirical methods (20). When a 6% solution of aqueous gelatin at pH 7.2 was treated with 2000 ppm 13CH2O for 24 h, the resulting ¹³C-NMR spectrum (Figure 1) of the hardened gel exhibited three new peaks: 74.1, 67.7, and 62.1 ppm (peaks C, D, and E, respectively, Figure 1). The first two peaks, C (74.1 ppm) and D (67.7 ppm), were assigned to methylols of the ε-amino and guanidino funtionalities of lysine and arginine (16-19), respectively, while the peak at 62.1 ppm (peak E, Figure 1) was assigned to the lysine-arginine methylene crosslink (16-19).
Figure 1. ¹³C-NMR spectrum of gelatin (6% in H2O/D2O) solution (1.0 ml) crosslinked for 24 h at pH 7.2 with 2000 ppm ¹³CH2O.
Incubation of 1.0 mL of the above solution with 0.05 mL of 3M HCl at 37°C for 24h effected the reversion of the gelatin solution to fluid state and resulted in the reduction in intensity of the lysine-arginine crosslink peak at 62.1 ppm in the ¹³C-NMR spectrum (peak E, Figure 2). When the pH of the aqueous gelatin solution was readjusted to 7.2 by the addition of 3M NaOH, not only did the neutralized solution return to gel state, but the ¹³C-NMR confirmed the reappearance of the lysine-arginine crosslink (peak E, Figure 3). The chemical structure proposed for the lysine-arginine crosslink in gelatin is the aminal function (1,16-19). Like the acetal, in which nitrogen atoms are replaced by oxygen atoms, the aminal would be expected to exhibit stability in basic media and susceptibility to acidic media (21).
It could be argued that the heat (37°C) employed in the incubatory period of the crosslinked gelatin was sufficient to hydrolyze peptide bonds, and that this event alone could explain the reduction in viscosity of the crosslinked gelatin after 24 h (22). However, the use of moderate temperature (37°C) and a short incubation time (24 h) in this experiment ensured minimal peptide bond hydrolysis. This notion was further supported by the fact that the when the crosslinked gelatin solution, incubated with acid (3M HCl, 0.05 mL) for 24 h at 37°C, was neutralized to pH 7.2, it returned to a viscous gel. Had there been significant peptide bond hydrolysis, the gelatin solution would have failed to gel, regardless of the quantity of crosslinking agent employed (23).
Figure 2. ¹³C-NMR spectrum of gelatin (6% in H2O/D2O) solution (1.0 ml) crosslinked for 24 h at pH 7.2 with 2000 ppm ¹³CH2O, after 24 h incubation with one drop 3M HCl.
Figure 3. ¹³C-NMR spectrum of gelatin (6% in H2O/D2O) solution (1.0 ml) crosslinked for 24 h at pH 7.2 with 2000 ppm ¹³CH2O, after 24 h incubation with one drop 3M HCl and neutralized to pH 7.2.
A second experiment was conducted, using 1.0 mL of a 6% aqueous gelatin gel solution crosslinked for 24 h at pH 7.2 with 2000 ppm ¹³CH2O. Incubation of the latter solution with 1.0 mg pepsin (sans acid) at 37°C for 24 h affected the reversion of the gelatin gel to fluid state, despite the fact that a ¹³C-NMR spectrum of the same solution confirmed the existence of 13C-formaldehyde-induced crosslinks between arginine and lysine residues (peak E, Figure 4). The resistance of the lysine-arginine crosslink to hydrolysis confirms enzyme-substrate specificity. Pepsin cleaves peptides at the carboxyl ends of hydrophobic and aromatic amino acid residues (e.g. tyrosine, phenylalanine, methionine) (24). The decrease in molecular weight of the partially digested gelatin molecule, induced by the presence of pepsin, affected a drop in viscosity of the aqueous gelatin solution. These results are analogous to previous studies (19), which established the resistance of the lysine-arginine (in an aqueous solution of formaldehyde-crosslinked gelatin) crosslink to pancreatic hydrolysis. This work also proposed that the decrease in viscosity of the crosslinked gelatin solution was due to pancreatic digestion of peptide bonds in the gelatin molecule, specifically ones which are on the carboxyl ends of lysine or arginine residues (19,24).
Figure 4. ¹³C-NMR spectrum of gelatin (6% in H2O/D2O) solution (1.0 ml) crosslinked for 24 h at pH 7.2 with 2000 ppm ¹³CH2O, after 24 h incubation with 1.0 mg pepsin.
Finally, when 1.0 mg pepsin and 0.05 mL 3M HCl was added to 1.0 mL of a 6% aqueous gelatin gel, crosslinked at pH 7.2 for 24 h with 2000 ppm ¹³CH2O, the hardened gel, after incubation for an additional 24 h at 37°C with enzyme and acid, reverted to fluid state. The ¹³C-NMR spectrum of the same solution, taken after the 24 h incubation at 37°C with enzyme and acid, verified a diminished intensity of lysine-arginine crosslink peak (peak E, Figure 5). In this experiment, however, when the pH of the gelatin solution containing pepsin and acid was taken to 7.2, the neutralized solution remained in the fluid state. A ¹³C-NMR spectrum was obtained of the neutralized gelatin solution (Figure 6), and the lysine-arginine crosslink peak (peak E, Figure 6) reappeared with an intensity comparable to the previous experiment involving a formaldehyde-crosslinked gelatin solution treated only with 3M HCl and neutralized (peak E, Figure 3). The major difference here is that while the 3M HCl alone caused a reversible hydrolysis (of the lysine-arginine crosslink) and viscosity in formaldehyde crosslinked gelatin (Figures 2 and 3), the inclusion of pepsin in the hardened matrix effected a permanent (irreversible) decrease in the viscosity (and molecular weight) of the same solution.
Figure 5. ¹³C-NMR spectrum of gelatin (6% in H2O/D2O) solution (1.0 ml) crosslinked for 24 h at pH 7.2 with 2000 ppm ¹³CH2O, after 24 h incubation with one drop 3M HCl and 1.0 mg pepsin
Figure 6. ¹³C-NMR spectrum of gelatin (6% in H2O/D2O) solution (1.0 ml) crosslinked for 24 h at pH 7.2 with 2000 ppm ¹³CH2O, after 24 h incubation with one drop 3M HCl and 1.0 mg pepsin and neutralized to pH 7.2.
As a control for the three previous experiments (crosslinked gel + acid, crosslinked gel + pepsin, or crosslinked gel + acid and pepsin), heat (37°C/24 h) was investigated with respect to its effect on both the intensity of the lysine-arginine crosslink and the physical state (fluid or gel) of the gelatin solution. Accordingly, the ¹³C-NMR spectrum of the crosslinked (2000 ppm ¹³CH2O/24 h) gelatin solution, after being heated in the absence of acid or pepsin for 24 h at 37°C, showed no detectable difference in the intensity of lysine-arginine crosslink peak as compared to the ¹³C-NMR spectrum from an unheated sample. Furthermore, both the heated (37°C/24 h; no acid or pepsin) and unheated gelatin solutions, which had been crosslinked with 2000 ppm ¹³CH2O for 24 h, remained in the gel state.
These experiments confirm not only the resistance of the lysine-arginine crosslink to hydrolysis by pepsin but also the susceptibility of the same crosslink to acid-mediated hydrolysis. The ¹³C-NMR spectra, which demonstrate the hydrolytic action of acid on the lysine-arginine crosslink (peak E, Figures 2,3,5,6), support the fact that its chemical structure within gelatin is an aminal functionality, involving a methylene bridge between two nitrogen atoms. Reduction in molecular weight of the gelatin polypeptide and the concomitant decrease in viscosity of the crosslinked gelatin solution was realized by incorporation of either acid or pepsin into the crosslinked gelatin solution. The chemical mechanism by which the aforementioned transformation was accomplished differed between the two agents. Specifically, pepsin was shown to depolymerize crosslinked gelatin while leaving intact the formaldehyde-induced arginine-lysine bridge. On the other hand, the effect of dilute acid on the arginine-lysine crosslink was clearly evident by the decreased intensity of the corresponding peak in the 13C-NMR spectrum (peak E, Figure 2).
Although both acid and pepsin treatments individually rendered a decrease in viscosity (gel to fluid state) of the crosslinked gelatin solutions, only the pepsin treatment effected a permanent fluid state. The crosslinked gelatin solution treated with acid, when neutralized to pH 7.2, easily reverted from fluid to gel state, and this event was accompanied by an increase in intensity of the arginine-lysine crosslink peak in the ¹³C-NMR spectrum (peak E, Figure 3). Confirmation of the reversible nature of the ¹³C-formaldehyde-induced crosslinking reaction was obtained on the hardened gel which was treated with both pepsin and acid, then neutralized to pH 7.2. The aforementioned gel contained a crosslinked, albeit fluid, gel.
To test for correlation between the presence of the arginine-lysine crosslink and pH within the formaldehyde-crosslinked gelatin solution, solutions of 6% gelatin, crosslinked for 24 h with 2000 ppm ¹³CH2O, were incubated for 24h at 37°C with varying amounts of 3M HCl (Table 2). Carbon-13 NMR spectra obtained for each solution, and the intensity of the arginine-lysine crosslink peak was plotted against the pH of the crosslinked gelatin solution. Regression of the intensity of the arginine-lysine crosslink in the aqueous gelatin solutions against the pH of the respective solutions demonstrated a linear correlation ((r2=0.91). This finding concurs with earlier work, which predicted that methylene-bis-amines (aminals) may be susceptible to hydrolytic cleavage by aqueous acids (21). The significance of the pH-crosslinking relationship is the following: (1) the severity of formaldehyde-induced crosslinking in gelatin may be reduced or even prevented by control of the pH of the gelatin solution during hard or soft capsule manufacture; (2) the impact of formaldehyde-induced crosslinking in gelatin capsules on the in vivo release of capsule content may be predicted by knowing the pH of the solution into which the crosslinked capsule is introduced. The first point above has limited application, because the commercial manufacture of gelatin batches is constrained within a relatively narrow pH range (4.5 to 6). Manufacture outside this window adversely affects physical properties such as viscosity, capsule brittleness, film strength, setting characteristics, and seam sealing (soft gelatin capsules). The second point is more practical: the effect of gelatin capsule crosslinking on in vivo performance may relate to the state (fed vs. fasted) of the upper gastrointestinal tract: residence time in a full stomach will be significantly higher than in an empty stomach (25). Thus, a crosslinked capsule may pass virtually intact through an empty stomach and release a negligible portion of its contents. In a full stomach, however, the same capsule may open due to longer exposure time in the acidic environment and greater extent of crosslink hydrolysis.
Because these crosslinking studies were limited to comparatively dilute (94% water, 6% gelatin) gelatin solutions compared to dry gelatin capsules (6-16% water), it could be argued that crosslinking may proceed more extensively in finished dry capsules than in the relatively dilute gelatin solutions utilized in these studies. Mass balance dictates that crosslinking processes which involve water as a product are favored at lower water concentrations. The potential consequence of the aforementioned is that pepsin and/or acid may not be capable of mitigating the insolubility of severely crosslinked gelatin in vitro or in vivo. The resolution of this dilemma could be achieved with an in vitro-in vivo γ-scintigraphic study using a γ-emitting isotope probe and an appropriate drug marker (26). The latter would be chosen based on ease of assay and relative freedom from hepatic metabolism. Stressed (crosslinked) and fresh capsules could be compared with respect to parameters such as bioequivalence (AUC, Tmax , Cmax ) as well as specific regions and times of capsule opening in fed and fasted human subjects (26).
Pepsin and acid depolymerize gelatin by unique mechanisms:
- The lysine – arginine crosslink in gelatin, crosslinked with formaldehyde, is resistant to hydrolysis by pepsin.
- The lysine – arginine crosslink in gelatin, crosslinked with formaldehyde, is proportional to the pH of the solution.