Chemically-modified polysaccharides for enzymatically-controlled oral drug delivery

Joseph Kost and Shmuel Shefer Depaftment of Chemical Engineering, Ben-Gurion University of the Negev, Beer-Sheva, 84 105, Israel

(Received 26 April 1990; accepted 7 June 1990)

Starch polysaccharides were investigated as bioerodible matrices for enzymatically-controlled oral drug delivery. Corn starch was ionically cross-linked with calcium chloride. It was found that the cross-linked starch could be used for entrapment and controlled release of bioactive molecules. The release rate of large molecules was degradation-dependent due to amylase activity, which might target the release to the intestine, when the particles were taken orally. The release of small molecules such as salicyclic acid, was mainly by diffusion.

Keywords: Drug delivery, polysaccharide, biodegradation, amylase

Matrix-controlled release devices are divided into two major

groups, erodible and non-erodible. The rationale for preferring

bioerodible or biodegradable systems is that such polymers

are eventually absorbed by the body and need not be

surgically removed. However, this advantage must be

weighed against the potential disadvantage that the

degradation products may be toxic.

In considering the properties desirable for a bioerodible

drug-carrier, starch polysaccharides were evaluated as drug-

delivery matrices because the degradation products of these

polymers occur naturally in the human body.

Starch granules are composed of two polysaccharides,

amylopectin and amylose; the former constitutes about 80%

of the most common starches. Amylose is essentially a linear

polymer having a MW of 100 000-500 000. Conversely,

amylopectin is a highly branched polymer with a molecular

weight in the millions. The branches of amylopectin contain

about 20-25 glucose units. Starch hydrolyses completely to

yield D-glucose. Amylases are capable of catalysing the

hydrolysis of a(D(l-4)) linkage in starch. In general,

a-amylases are endoglycosidases, attacking glucan some-

where away from the chain ends at an internal glycosidic

bond, producing varying types of oligosaccharides,

/Gamylases attack glucan from the non-reducing end to

produce maltose and glucoamylases degrade starch

molecules to glucose. a-amylases are the most common. In

humans, they are produced in the salivary glands and the

pancreas.

Starch, in its native or modified form, is used extensively

in the food industry as a carbohydrate source, extender,

processing aid, thickener, stabilizer and texture modifier’. It

is also used as a skin emollient and as an antidote for iodine

Correspondence to Dr J. Kost.

0 1990 Butterworth-He~nemann Ltd. 0142-9612/90/090695-04

poisoning. Other pharmaceutical applications include its use

as a tablet filler, binder and disintegrant’.

Starch cross-linking is widely used to provide textural

characteristics necessary in food systems. The basic idea

behind cross-linking is the toughening of starch granules by

treatment with di- or polyfunctional reagents capable of

reacting with the hydroxyl groups in the starch molecules.

Calcium may form ionic cross-links between the hydroxylic

groups on the polysaccharide molecules as presented in the

following scheme:

2 -O,,NdOH_2 -0 Nat “” i i -j- ! 0 'Ca' 0- '+NNa'

gelatln,ratloll

Recently, several reports have described methods for

encapsulating herbicides and pesticides in starch

matrices3“. This paper outlines work on evaluation and

characterization of ionically cross-linked corn starch for

enzymatically-controlled oral drug delivery.

MATERIALS AND METHODS

Corn starch granules (Sigma 4126) were ionically cross-

linked by calcium chloride using a modification of the

procedure developed by Trimnell et a/.6. To study the release

mechanism, the granules were loaded with low- or high-

molecular weight model drugs, such as sodium salicyclic

acid (SA) (Sigma 3007). myoglobin (Sigma 1882) or bovine

serum albumin (EGA) (Sigma 4503). Sodium hydroxide and

calcium chloride were reagent grade.

In the processing procedure, 12 g starch with the

releasing agent was suspended in 30 ml water and 50 ml,

6.6 wt% sodium hydroxide solution was added. After 1 h of

mixing at 600 rev min-‘, a solution of CaCI, (0.5 g/ml) was

Biomatenals 1990, Vol 11 November 695

Polysaccharides for enzymatically controlled delivery: J. Kost and S. Shefer

added, which coagulated the pastes upon mixing. The resulting matrices were air dried and ground in a Waring blender. The products were sieved to specific particles size range (0.6-I mm).

As each of the starch monomeric units contained three hydroxyl groups, the weight of calcium chloride added was calculated so that the amount of bivalent calcium ions added would be equivalent to twice the number of monovalent hydroxylic groups on the starch. This concentration of calcium ions was defined as a theoretical degree of substitution DS = 3 (2/3 of that calcium amount was defined as DS = 2, etc.).

Matrix degradation and drug release were detected in simulated stomach and intestine solutions in the presence of a-amylase. The release experiments were performed in mixed vessels (rotating shaker, 110 rev min-‘) at 37”C, containing simulated gastric (hydrochloric buffer, 0.1 M,

pH 2) or intestine (phosphate buffer, 0.1 M, pH 7.4) solutions, and a-amylase (EC 3.2.1 .I type 2A. from bacillus species, Sigma 6380). Samples were analysed spectrophoto- metrically (Milton Roy, Spectronic 1001). Amylase activity was determined by maltose accumulation*, using a reducing sugars assay with 3,5-dinitrosalicylic acid reagent (Sigma 15 IO)’ at 540 nm. Release rates of SA, BSAand myoglobin were followed by monitoring the absorbance at 260, 220 and 4 10 nm, respectively.

The starch granule surfaces were examined in a scanning electron microscope (SEM) (Jeol, JSM-35CF). which was equipped with an analyser to record the elemental distribution and element count.

RESULTS AND DISCUSSION

The formation of calcium cross-links should tighten the starch matrix and therefore affect both its degradation rate and the diff usion of molecules through the matrix. The extent of cross-linking as a function of calcium concentration added was examined, evaluating the elemental distribution and elemental counts on the surfaces of rinsed starch granules. The decrease in the elemental ratio of chloride to calcium (CVCa) with the concentration of CaCI, added (Figure 7) suggested that calcium was incorporated into the matrix by the ionic interactions, whilst the non-interacting chloride was removed from the surface when rinsed. The linear relation between CVCa ratio and DS suggested that in this range of (calcium chloride)/starch concentrations, there was a linear proportion between the amount of calcium chloride added and the cross-linking density. In control corn starch samples not activated by sodium hydroxide, or polymers

0.14,

0.12

I

G L 0.10 - 0

0.08 -

“.“”

0 1 2 3 4

DEGREE OF SUBSTITUTION

Figure 1 Chloride to calcium ratio as detected by an elemental analyser on modified starch surfaces processed with calcium chloride, in amounts theoretically sufficient for degree of substitution DS = 1, 2 or 3 hydroxyl groups on each starch monomeric unit.

without hydroxyl groups (dialdehyde starch, methyl cellulose), no effect of the amount of calcium chloride added on the ratio of CI/Ca could be detected.

The enzymatic activity of a-amylase significantly increased the degradation rate of the modified starch (Figure 2). In the absence of amylase or in simulated gastric (acidic) solution (Figure 3), where the enzymatic activity was significantly lower, the degradation rate of the modified starch granules was very low.

The effect of a-amylase activity on surfaces of modified starch granules was visualized on SEM micro- graphs, comparing surfaces exposed to buffer solution (pH 7.4) without amylase enzymes, to surfaces exposed to solution (pH 7.4) containing a-amylase (Figures 4 and 5). The enzymes degraded the starch, creating a very porous structure, whilst the surfaces of the control starch granules not exposed to a-amylase were relatively smooth.

The release studies demonstrated that the release rates of high-molecular weight BSA molecules were low in the stomach environment (Figure 6) but higher in the intestine (Figure 7) and closely related to the matrix degradation rates (Figure 2). The release rates of the low- molecular weight SA from the cross-linked starch matrices were higher than those of large molecules. The release rates of small molecules were not affected by amylase activity, suggesting thatthe release of small molecules was mainly by diffusion.

Figure 8 shows that the release rates of the high- molecular weight substances such as BSA (68 000) and

0.8 1 T

0.6 - 8

i-

d 0.4 -

o.o; 10 20 30 40 50 60

TIME (hr)

Figure 2 Calcium-modified corn starch (DS = 3) fraction degraded in simulated intestine solution versus time, presented as the fraction of maltose accumulated divided by the initial weight of the starch sample (maltose,), in the presence (0) and absence (A) of 0.5 unit/ml of a- amylase. Vertical bars indicate SD (n = 12).

0.010

0.005

0.000 4 6 8

TIME (hr)

Figure 3 Calcium-modified corn starch (DS = 3J fraction degraded in simulated stomach solution versus time, presented as the fraction of maltose accumulated, divided by the initial weight of the starch sample (maltose,), in the presence (0) and absence (A) of 0.5 unit/ml of a- amylase. Vertical bars indicate SD fn = 3).

696 Biomaterials 1990, Vol 11 November

Polysaccharides for enzymatically controlled delivery: J. Kost and S. Shefer

Figure 4 Scanning electron micrograph of calcium-modified corn starch granule surface (OS = 3) exposed for 5 h to phosphate buffer at pH 7.4. without enzyme.

Figure 5 Scanning electron micrograph of calcium-modified corn starch granule surface (DS = 3) exposed for5 h to phosphate bufferatpH 7.4, with 0.5 unit/ml of a-amylase enzyme.

0 1 2 3 4 5 6

TIME (hr)

Figure 6 Fraction of BSA released versus time from calcium-modified corn starch (DS = 3, 1 wt% BSA) in simulated stomach solution (pH 2). in the presence (0) and absence (A) of 0.5 unit/ml of a-amylase enzyme. Vertical bars indicate SD /n = 3).

myoglobin (17 800) in the absence of amylase were lower than in its presence (Figure 9). In the absence of amylase, the smaller molecule, myoglobin, was released faster, suggesting that the release mechanism in the absence of enzyme for large and small molecules (Figure 7) was diffusion- dependent. In contrast, in the intestine environment in the presence of a-amylase, where the release of large molecules was degradation-dependent, there was no effect of the size and BSA and myoglobin were released at the same rate (Figure 9).

0.8

e cn a

y 0.6 3

0 2 4 6 8 TIME (hr)

Figure 7 SA and BSA fraction released from calcium-modified corn starch (DS = 3. 1 wt% BSA or SA), in simulated intestine solution (pH 7.4). in the presence (+EJ and absence I-E) of 0.5 unit/ml of a-amylase. Vertical bars indicate SD (n = 3). (A) SA + E, (A) SA - E. (*J BSA t E, (0) BSA - E.

6

TIME (hr)

Figure B Myoglobin (A) and BSA (0) fraction released from caloum- modified corn starch (DS = 3. 4 wt% myoglobin or BSAJ. in simulated intestine solution in the absence of enzyme. Vertical bars indicate SD

(n = 3).

0.5, I

__” - -

0 1 2 4 5 6

TIME (hr)

Figure 9 Myoglobin (A) and BSA (0) fraction released from corn starch (DS = 3, 4 wt% myoglobin or BSA) in simulated intestine solution with 0.5 unit/ml of a-amylase. Vertical bars indicate SD (n = 3).

Eiomaterials 1990, Vol 11 November 697

Polysaccharidas for enzymatically controlled delivery: J. Kost and S. Shafer

CONCLUSIONS

Starch granules could be used for entrapment and controlled release of bioactive molecules when cross-linked by calcium. The release rate of high-molecular weight molecules was degradation-dependent, mainly due to amylase activity, targeting the release to organs which had amylase activity: saliva, intestine and blood. Small molecules were released by diffusion. The results suggested possible advantages of starch matrices fororal delivery systems: (1) the matrices are approved as food additives, (2) their degradation products occur naturally in the human body, and (3) the enzymatically- controlled drug delivery may target the release to the intestine and eliminate the need for enteric coating.

REFERENCES 1 Wurzburg. O.B., Starch in the food industry, in Handbook of Food

Additives, (Ed. T.E. Furia) Boca Raton, FL, USA, CRC press, 1968: 377

2 Franz, G., Polysaccharides in pharmacy, in Pharmacy, Thermo- mechanics, Elastomers, Telechelics, (Ed. K. Dusek) Berlin, Germany, Springer-Verlag, 1986; l-30

3 Shasha, B.S., Trimnell, D. and Otey, F.H., Encapsulation of pesticides

in a starch-calcium adduct. J. Polym. Sci., Polym. Chem. 1981, 19, 1891-1899

4 Trimnell. D., Shasha, B.S., Wing, R.E. and Otey. F.H.. Pesticide encapsulation using a starch-borate complex as wall matrix, J. Appl. Polym. Sci., 1982, 27, 3919-3928

5 Shasha, B.S., Trimnell, D. and Otey, F.H., Starch-borate complex for EPTC encapsulation, J. Appl. Polym. Sci., 1984, 29, 67-74

6 Trimnell. D., Shasha, B.S. and Otey. F.H., The effect of alfa amylase upon release of trifluralin encapsulated in starch, J. Control. Rel. 1985, 1, 183-190

7 Wing, R.E.. Maiti. S. and Doane, W.M., Factors affecting release of butylate from calcium ion modified starch-borate matrix, J. Control. Rel. 1987, 5, 79-89

8 Bernfeld. P.. Amylases alfa and beta, Methods in Enzymol. 1955, 1, 149-158

9 Hostettler, F., Borel, E. and Duel, H.. Reduction of 3,5-dinrtrosalycilrc acid by sugars, He/v. Chim. Acta. 1951, 34, 2132-2139

698 Biomaterials 1990, Vol 11 November