726
R.W. Kalicharan et al. / International Journal of Pharmaceutics 515 (2016) 721–728
*
In literature, a half-life time hydrolysis of 4 min for 2.5
This suggests that there must be another process that plays a role.
This process is the diffusion of ND through the cell membrane
which is the rate-limiting step. Otherwise, Michaelis-Menten
kinetics would have been observed and a lag time absent. As ND is
more lipophilic (log P = 8.1 (“ChemSpider,” 2015)) than N (log P = 3.0
(“ChemSpider,” 2015)), we assume that the limiting step in mass
transfer is the cell influx of ND rather than the cell efflux of N.
10ꢄ3
m
mol/mL (1
m
g/mL) nandrolone phenyl propionate in rat
plasma has been reported (14). Based on these data, the duration
* 10ꢄ5
time to hydrolyse 2.3
mmol/mL (=0.01 mg/mL) ND in HS was
proposed by Wijnand et al. (Wijnand et al., 1985) to be below one
hour. However, as reported by Li et al. and Rudakova et al. (Li et al.,
2005; Rudakova et al., 2011), no carboxylesterases are present in
HP. This implies that no enzymatical hydrolysis of ND to N can
occur in HP and HS, which is confirmed in this current study: In
human serum and plasma, no hydrolysis of ND occurred during 5 h
of incubation. Also, no chemical hydrolysis of ND occurred in these
media. Moreover, this entails that ND hydrolysis does not occur in
interstitial fluid since interstitial fluid originates from blood
plasma. It contains water and molecules less than approximately
40 kD molecular weight (Charman and Stella, 1992; Wiig et al.,
2012). This means that even when carboxylesterases would be
present in human blood plasma, they would not appear in the
interstitial fluid as the molecular weight of human carboxylester-
ases is approximately 60 kD (Imai and Ohura, 2010).
This paper demonstrates that ND hydrolysis does occur in
human blood. Apparently human blood contains carboxylester-
ases. The appearance of N was observed after an average lag time of
34.9 ꢀ 2.5 min. Noticeable was the absence of Michaelis-Menten
behaviour. Instead of nonlinear enzyme kinetics, a linear rate of
hydrolysis was seen (Fig. 4). These observations suggest that the
carboxylesterases, involved in ND hydrolysis, must be present in
the blood cells since the distinction between whole blood and
plasma is the presence of cells. Hydrolysis can occur either on the
blood cell membrane, or occurs intracellularly. Because a lag time
of N appearance was noticed, carboxylesterase activity on the cell
membrane is apparently not very pronounced. Otherwise, the N
recovery should be seen instantaneously. It can therefore be
concluded that hydrolysis occurs after membrane diffusion and,
subsequently intracellularly in erythrocytes cytosol (Quon and
Stampfli, 1985) and probably in leukocytes. This is in line with
literature, where it is reported that carboxylesterases are present
in the endoplasmic reticulum and cytosol in human cells.
In porcine muscle and subcutaneous tissues, little hydrolysis of
ND occurred in porcine muscle tissue during 5 h of incubation
(Fig. 5). The tissues were chosen to be originated from a pig instead
of other species (such as rats) to prevent false-positive results.
Similar to human tissues, interstitial fluid between porcine tissue
cells does not contain carboxylesterases (Li et al., 2005). Because
the penetration depth of lipophilic molecules into tissue is very
low (Lerner et al., 2006), it is relevant to study the carboxylesterase
activity in cells that are in the direct proximity of the oil depot. The
results with tissue conclude that hydrolysis in porcine muscle and
subcutaneous cells occurs slowly, indicating that only little ND has
reached an enzyme. In line with hydrolysis in HB, this also suggests
that in vivo hydrolysis only occurs intracellularly, because
immediate (within 30 min) N appearance was absent.
Therefore, the
N efflux will be neglected in the following
estimation. With the value of lag time being roughly 2100 s and
the literature value of the erythrocyte membrane thickness (h)
(7*10ꢄ7 cm (Changjun Liu et al., 2003)), the diffusion coefficient (D)
of ND through this membrane can be estimated to be
3.9*10ꢄ17 cm2/s according to the following equation:
h2
Lagtime ¼
6D
This calculated value of the diffusion coefficient is much lower
than other steroid diffusion coefficients. For example, the D of
testosterone in percutaneous absorption has been reported to be
1.95*10ꢄ11 cm2/s (Scheuplein et al., 1969). Calculated lag time for
this testosterone would be 4.2*10ꢄ3 s (h was kept constant), which
gives a negligible corrected (for the N efflux) diffusion coefficient
for ND in current study (D remains 3.9*10ꢄ17 cm2/s). There is a
significant difference between the values of ND and testosterone
and this may be a source of some discussion. However, it is at least
a reflection of the large difference that exists between a prodrug
having a log P = 8 and a parent drug having a log P = 3. The
conclusion of this observation is that ND hardly does permeate
through tissue. This indicates that ND must be absorbed into the
central circulation via other routes, as direct absorption is excluded
due to the high partition coefficient. It is assumed that ND
subsequently adheres to small proteins (<40 kD) and migrates into
lymph vessels to be absorbed into the central circulation.
Recently, it was shown that the in situ surface of a 0.5 mL
injected oil depot is 755.4 mm2 (Kalicharan et al., 2016a). As can be
seen in Table 2, the amount of cells per 1000 mm2 tissue is around
105–108 cells, which is negligible with the amount of cells in blood
(1015 cells/mL). Another advantage of blood, is that it is continu-
ously refreshed (sink conditions).
Benzyl alcohol (BOH) is commonly used in oil depots at a
concentration of 1–10% (m/v) (Bagchus et al., 2005; Kalicharan
et al., 2016c; Minto et al., 1997; Van Weringh et al., 1994; Wijnand
et al., 1985). It is used as viscosity reducer, local anaesthetic and as
co-solvent in oil depots (Rowe et al., 2009). Recently, BOH was
reported to have a very different absorption profile than N
(Kalicharan et al., 2016c). In contrast to N, BOH was detected in the
bloodstream within minutes after injection. Furthermore, whereas
N was measured for 5 weeks after injection, BOH was cleared form
the central circulation within 36 h (Kalicharan et al., 2016b).
The BOH molecular structure shows great similarity to the one
of benzil (Table 3), which is an inhibitor of carboxylesterases
(Hatfield and Potter, 2011). The inhibition is due to steric
interaction of the benzene ring of benzil with the pocket of
human carboxylesterase 1 (hCE1) (Harada et al., 2009). This was
Before intracellular hydrolysis can occur, ND must diffuse
through the cell membrane. As can be seen in Fig. 4, the enzymatic
conversion is proportional to the concentration of the substrate.
Table 2
Estimation of amount of cells in studied media nandrolone decanoate may encounter during absorption into the central circulation.
Type
Amount of cells
Amount of cells available for ND at tissue surfacec
Subcutaneous cellsa
2.6 * 103 cells/mg tissue
9.3 * 104 cells/mg tissue
4.2–5.5 * 1015 cells/ml blood
1.4 * 102 cells/mm2
3.2 * 105 cells/mm2
–
Muscle cellsb
Erythrocytes
a
The volume of human adipose cells is 660 *10ꢄ6 mm3 (26). The 590 mg porcine subcutaneous tissue had a volume of approximately 1 cm3 (=1000 mm3).
No dimensions of human muscle cells are published. The following data is used: rat myocytes have a volume of 15.6 * 10ꢄ6 mm3 (27). The 690 mg porcine muscle tissue
b
had a volume of approximately 1 cm3 (=1000 mm3).
c
the diameter of human subcutaneous cells and muscle cells are respectively 94 and 2 mm (28,29).
Kalicharan
