Enhanced accumulation of ribavirin and its ...

420

lTAL J GASTROENTEROL HEPATOL 1997:29:420-6

Enhanced accumulation of ribavirin and its metabolites in liver versus erythrocytes in mice administered with the liver targeted drug G. Di Stefanol, A. Bignaminiz, C. Busil, F.P. Colonna3, L. Fiume'

Background. Ribavirin increases the eficacy of a-interferon in chronic hepatitis C, but accumulates in erythrocytes causing haemolysis. Aims. To reduce this side effect we conjugated ribavirin with lactosaminated poly-L-lysine. Methods. In mice administered with the same dose offree o r conjugated [3H]ribavirin we determined the levels of radioactivity in liver and erythrocytes and measured the hepatic concentrations of ribavirin triphosphate. Moreover; we determined the doses o f f r e e and conjugated ribavirin producingu a 50% reduction in the virus titre IEDro) , --,in liver o f mice infected with murine hepatitis virus. Results. In mice treated with the conjugate, the ratio dpm in liver/dpm in erythrocytes was 2 . 2 or 4.7-fold higher than in animals administered with the free drug given intramuscularly o r orally, respectively. The concentration of [-'H]ribavirin triphosphate was found to be 2-fold higher in mice injected with the conjugated drug than in animals orally treated with free ribavirin. In murine hepatitis virus infected mice, the EDJOwas 27.4 pg/g for conjugated ribavirin and 47.2 pg/g for thefcee drug. Conclusions. These results support the possibilify that coniu.. gated ribavirin may produce the same pharmacological activity in liver a s thefree drug but with a reduced haemolysis. Ital J Gastroenterol Hepatol 1997;29:420-6 ~ Key words: Asialoglycoprotein receptor - D N targeting - Hepatitis C - Ribavirin

Introduction Ribavirin (RIBV) is a broad spectrum antiviral nucleo..

of Experimental Pathology. Universiry of Bologna, Bologna; HyperPhaq SpA, Milan; ' 1.Co.CE.A. Institute. CNR, Bologna, Italy.

'

Aclolowledgements: This work war supponed by gronrs fmm AIRC, the 1.stiruto Superiore di Sanirci (Pmgelto Epafite Kmle), MURST and the Urtiversity of Bologno (finds for selccted msearch topics). Address for correspondence: Proj L. Fiume, Dipanimento di Patologia Sperimentale. via San Giacomo 14, 40126 Bologna. Italy - Far: +39-51-354746. Manuscript submilred July 11. 1997 and accepted in rwised fonn September 17, 1997.

side analogue ' which, when administered to patients with chronic hepatitis C in association with a-interferon (IFN), produces a sustained biochemical and virological response in a significantly higher number of patients than in the group treated with IFN alone ". The mechanism of the beneficial effect of RIBV in this infection is not known. A drawback of RIBV is its accumulation in red blood cells (RBC) which produces haemolysis 2-8. This side effect causes symptomatic anaemia in some patients and constantlv raises the iron concentration within hepatocytes to levels l o which were found to reduce the response to IFN 12. Drug accumulation in RBC and haemolysis are observed only in primates and do not occur in rodents '13. We considered the possibility that a liver targeting of RIBV could lower the haemolysis, thus enhancing the therapeutic effects of the IFN and RIBV association. We coupled RIBV to lactosaminated poly-L-lysine (LPoly(LYS)) 14. With the use of this galactosyl terminating hepatotropic canier, it has been possible to prepare conjugates for injection via the intramuscular (im) route which have displayed very good tolerability in acute and subchronic toxicity tests I4.l8. The L-Poly(LYS) conjugate of RIBV selectively penetrated mouse liver cells and incubated in vitro in human or mouse blood, did not enter into RBC. However, in mice administered with the conjugate, RIBV and its metabolites, after the intracellular release from the carrier, partly went out from hepatocytes and penetrated RBC 14. In the experiments reported here, we administered mice with the same dose of free or conjugated [3H]RIBVand measured how many times higher were the ratios between the levels of radioactivity in the liver and RBC in animals treated with the conjugate. The free drug was given by the irn route or by oral gavage (po); the conjugated drug was injected irn. In order to establish whether the radioactivity in liver of conjugate-treated mice was associated with pharmacologically active metabolites of RIBV, we compared the hepatic concentrations of ["IRIBV triphosphate in mice administered with the same dose of free or conjugated drug. Moreover, we determined the irn doses of free and conjugated RIBV which produced a 50% reduction of virus titre in the liver of mice infected with murine hepatitis virus (MHV).

"

Material and Methods Labelling of RIBV RIBV was obtained from Alfa-Wasserman (Bologna, Italy). Since the commercially available radioactive RIBV completely loses tritium during conjugation with L-Poly(LYS), we introduced tritium into the RIBV molecule by a stable bond. First, we oxidized the primary OH group of the sugar moiety of the drug and then reduced the nearly formed aldehyde with trtiated borohydride. RIBV was converted to the 5'-0dimethoxytrityl derivative; the 2'-3' OH were then protected using the t-butyldimethylsilyl group 19, which is stable during the oxidation and reduction steps and can be easily removed with tetrabutylammonium fluoride on SiOz (Fluka, Buchs, Switzerland). The primary OH group, deprotected by 80% acetic acid, was oxidized using pyridine sulphotrioxide in dimethylsulphoxide 202'. Then 1 mmol of 2'-3' protected RIBV was dissolved in 4 ml of well dried dirnethylsulphoxideldichloromethane 1:2; 4 equivalents of triethylamine and 2.1 equivalents of pyridine sulphotrioxide were added and the mixture stirred for 3 h under argon flushing at room temperature. The reaction was monitored by TLC eluted with 10% methanol in dichloromethane and sprayed with a 2-4 dinitrophenylhydrazine solution 22; the final product was characterized as a 1-2 dianilinoethane crystalline derivative 23 by 'H-NMR spectrum. The aldehyde was reduced by sodium [3H]-borohydride in phase transfer conditions 24. 240 mg protected RIBV 5'-aldehyde (0.5 mmol) in 4 ml dichloromethane was added to a water solution (0.6 ml) of 10 mCi NaB3H4 and 7.4 mg tetrabutylammonium bromide (0.024 mmol); the mixture was stirred for 2 h at room temperature. Then 23 mg NaBH4 (0.6 mmol) was added to complete the reduction and left for 2 h. The mixture was transferred using dichloromethane (2 x 6 ml) in a separatory funnel and washed with 4 ml of water. The water layer was extracted twice with dichloromethane (4 ml); the organic extract was dried over sodium sulphate and evaporated. Dry tetrahydrofurane (6 ml) was then added together with 1.1 g of tetrabutylammoniumfluoride on Si02(1.32 mmol) and the mixture was stirred for 2 h. The Si02 was filtered off and washed with water. Tetrahydrofurane was evaporated under vacuum and the product extracted with diethyl ether to remove the detached silyl groups. The water solution was then treated for 30 min with 400 rng of Dowex SOWX-H+ resin (Fluka) to remove tetrabutylammonium, filtered and lyophilized. The preparation of vitiated RIBV was performed twice, with a yield of about 30% each time. The purity of the two samples, assessed by high performance liquid chromatography (HPLC), were 70 and 80%; the specific activities (dpmlpg) of ['HIRTBV

were 300,000 and 150,000. ['HIRIBV was diluted with the unlabelled compound; the final preparations had a purity >95% and specific activities of 49,000 and 23,000 dpmlpg, respectively.

RIBV conjugation Conjugation of tritiated and unlabelled RIBV was performed via the imidazolide of its 5'-monophosphate ester (RIBVMP) 2s, as described by Di Stefano et al. 14. RIBVMP was coupled to a L-Poly(LYS) preparation in which 30-33% of lysine &-aminogroups had been substituted with lactose residues by reductive amination 26; the poly-L-lysine - HBr used in these experiments was obtained from Sigma (St. Louis, MO, USA) and had an average molecular weight of 30,000 Da. In the conjugates, RIBV content, determined according to Di Stefano et al. 14,ranged from 315 and 320 pglmg. Mice and drug administration Female Swiss CD-1 weighing 28-30 g and female 6week-old BalbIC mice, weighing 16-18 g were used. They were obtained from Harlan Nossan (Milano, Italy) and were maintained in an animal facility at the Department of Experimental Pathology, Bologna. They received humane care in accordance with the guidelines of the Italian Minishy of Health. Swiss CD-1 mice were used in the experiments in which total liver and RBC radioactivity was counted or hepatic [3H]RIBVTP concentration was measured; BalbIC mice were used in the antiviral study. Injections irn of free and conjugated [3H]RIBV were performed into the back muscles of the hind legs by using a 25 PIHamilton microsyringe. Compounds (7 pglg free or coupled ['HIRIBV) were administered in a total volume of 10 p1 of NaCl0.9%. When given by oral gavage, free ['HIRIBV (7 pglg) was dissolved in a volume of 100 pl of NaCl0.9% and mice were fasted 4 h before and 2 h after drug administration. Determination of radioactivity in liver and RBC At different times after free or coupled ["HIRIBV administration, mice were anaesthetized with ether and, after bleeding from the retro-orbital plexus, liver was rapidly removed and homogenized in 4 vol (wtlvol) of water. Mice died while under anaesthesia. For each compound and time interval, 2 to 4 mice were used. Haematocrit was determined and radioactivity of liver homogeuate, plasma and whole blood was measured. Liver homogenate (0.2 ml) and plasma (0.1 ml) were placed in counting vials and mixed with 1 ml soluene350 (Packard, Downers Grove, IL,USA). After 45 min. incubation at 50°C, 2-propanol (0.3 ml) and 30% hydrogen peroxide (0.3 ml) were added; the samples were counted after disappearance of foaming and addition of 10 ml Hionic Fluor (Packard). Whole blood (0.1 ml)

Liver targeted ribavirin

was placed in a glass counting vial; 1 M NaOH (0.5 ml), 2-propanol (0.5 ml) and 30% hydrogen peroxide (1.5 ml) were added. After standing at room temperature overnight, 20 ml of Hionic Fluor were added and radioactivity was counted. dpm/ml RBC were calculated from radioactivity of whole blood and of plasma taking into account the value of haematocrit '. Determination of hepatic kiH]RIBVTP The pharmacologically active metabolites of RIBV are the 5' mono-, di- and triphosphate derivatives ". We measured the triphosphate (13H]RIBVTP) since this compound can be easily identified in the chromatogram of liver extrace. Mice were killed under ether anaesthesia and 1 g liver was rapidly weighed and homogenized in 4 vol cold 5% perchloric acid. For each time and each compound, two animals were used and liver homogenates were pooled. After centrifugation, the snpernatants were neutralized with KOH. KC104 was removed and liver extracts were lyophilized. The dry material was dissolved with 2 ml water and clarified by centrifugation; 200 fl were analyzed by HPLC on an anionic exchange radial-pak cartridge 8PSAX (Waters, Millipore Corp., Milford, MA, USA) eluted with a linear gradient of KH2P04 from 20 mM to 1 M, pH 3.5, with a flow rate of 0.5 ml/min. A RIBVTP standard, synthesized according to Mishra and Bloom 2s, was added to each sample of liver extract; the amount of PHIRIBVTP was determined by counting the radioactivity eluting in the same position as the standard. Study of antiviral activity Antiviral activity was studied using the Friend-Braunsteiner strain of MHV, which was found to be sensitive to RIBV by Sidwell et al. 29. The virus was obtained from the American Type Culture Collection (ATCC, Rockville, MO, USA.). It was titrated on NCTC clone 1469 cells, as described below. BalbIC mice were inoculated ip with approximately 100 CLDso (lethal dose for 50% cultures in our titration assay). Free and conjugated RIBV were administered im twice daily beginning 2 h before virus inoculation. Ten animals were used in each treatment group. Mice were killed 48 h after virus inoculation. Titration was performed on monolayers of NCTC clone 1469 cells, grown in 96-well Falcon Microtest I11 tissue culture plates (Becton Dickinson, Lincoln Park, N.J., USA). Liver homogenates were diluted with MEM ten-fold through and 50 p l were added to the cell monolayer. Six wells were prepared for each dilution. After 2 h, the inoculum was removed and the cultures were maintained for three days or more in NCTC 135 medium supplemented with 2% chicken serum (Sigma). The viral cytopathic effect (complete destru~tion of the cell monolayer) was determined after 72 h. In cultures surviving this time, no cytopathic effect ap-

peared in the following days. Fifty-percent endpoints were determined by the method of Reed and Muench 30. Statistical andphannacokinetic analyses Radioactivity levels in liver and RBC as well as hepatic concentrations of [3H]RIBVTP have been examined as trapezoidal (model independent) AUC (area under cuwe) using the Siphar programme (Simed, Cr&teil, France). No modeling was applied since each monitored data point (dpm level or [3H]RIBVTPconcentration as a function of dmg form, route and time) was supplied by different animals (2-4 animals at each point). ANOVA was applied to the cornparison of the antiviral effect, using drug form and dose as independent factors. Dose-effect regression was examined with the multiple regression procedure. These analyses were performed with the SPSS programme (SPSS Inc., USA) using all individual data available.

Results Chemical characteristics of conjugate In the different conjugate preparations, RIBV content ranged from 315 to 320 pglmg. Previously, we reported that the solubility of coupled RIBV was 64 mglml (=200 mglml of conjugate) 14. We have now observed that by adding a dose of D(-)ribose equimolar to RIBV, the coupled drug can be dissolved to 150 mg/ml(=450

0

2

4

6

8

24

hours Fig. 1. Percentages of radioactivity remaining in the injected muscles after irn administration of free (U) and conjugated ['HIRTBV

(4(7 LIdg).

mg/ml of conjugate) maintaining a good fluidity. Moreover, ribose allows the conjugate to easily dissolve after lyophilization. The possible mechanism of the effect of ribose on conjugate solubility will be reported elsewhere.

I

Radioactivity distribution in liver and red blood cells Figure 1 shows the disappearance from the site of injection of free and conjugated ['HIRlBV administered im. Due to the much higher molecular weight, the absorption of the latter compound was slower; about 10% of the injected dose was still found after 6 h. However, at 24 h there was not a substantial difference in the percentages of injected radioactivity (about 2%) which remained in situ after administration of the free and coujugated drug. The levels of radioactivity in liver and RBC after administration of the labelled compounds are reported in Figure 2. Figure 3 shows the ratios dpm in liverldpm in RBC in animals administered with the conjugate or with the free drug. In mice treated with the conjugate, the trapezoidal AUC (0-24 h) of this ratio was 2.2 or 4.7 times higher than that in animals which received the free drug given im or po, respectively. The nmol of rH]RIBVTP per g liver are reported in Figure 4. The AUCs (0-24 h) of this metabolite were found to be 2-fold higher in mice injected with conju-

hours

hours

Fig. 2. Radioactivity in Liver (frame A) and RBC (frame B) of mice

administered conjugated ['HIRIBV (A) and free ['HIRIBV given im ( 0 ) or po (+).The earliest determinations were performed at 5 min (free [%I]RIBV, im), 15 min (free [WIRIBV, po) and I h (conjugated ['HIRIBV). Each envy represents the mean value of results from 2-4 animals. Livers were processed separately (standard errors ranged from 1 to 8% of mean values) and blood was pooled. The AUC (0-24 h) of the values of radioactivity in liver (dpm.h-S.A.-'.g~')were 158.29 (RBV, po), 213.10 (RIBV, im) and 393.33 (conjugated RIBV). The AUC (0-24 h) of the values of radioactivity in RBC (dpm.h.S.A:'.ml-I) were 36.01 (RIBV, po), 20.19 (RIBV, im) and 15.86 (conjugated RIBV).

!'.-:i>s dpm in livcr Jprn III KBC after administration of fnw I'IIIRIRV given po (*)or 1111 (01 imrl :nn~ugatcd ('HIRIRV (A) Ratios were obtained from data in Figure 2. The AUC (0-24 h) were 114.20 (RIBV, po), 241.15 (RIBV, im) and 539.88 (conjugated RIBV). Fig. ?.

424

Liver targeted ribsvirin

loo T

Free RlBV

0

10

20

30

40

dose (pglglday) 0

8

12

18

24

hours Fig. 4. Concentrations of ['HIRIBVTP in mouse liver after administration of free [WNIBV given po (+) or im (0)and of conjurated 13H1RIBV(A). . , The earliest determinations were ~erformed at Ihr s m e t~mesac i n Fleurc ?. 'I'he AUC (0-24 h, (nrnol I , $ I w ~ . r ~ . 6 8 . 0(RIBV. 4 PO,. 1 3 1 77 (RIRV, im) and 136.54 ( c ( I ~ ~ u ~ . I I c J RIBV).

-

I

I

Conjugated RlBV

. .

gated RIBV than in animals treated po with the same dose of the free drug. Antiviral activity

Figure 5 shows the relation between the virus titre in liver and the daily doses of free and conjugated RIBV administered to mice. The regression equations of titre vs dose, using all available individual data in the estimation, is: titre=5.55 (-c 0.22) - 0.101 (* 0.021) x dose with conjugated RIBV, and titre=5.65 (+ 0.22) 0.061 (k 0.010) x dose with free RIBV. The estimation of the dose yielding a 50% decrease of virus titre yields, respectively, 27.4 and 47.2 pglg. Thus, the antiviral potency of conjugated versus free RIBV, as either EDSOor slope ratio, is 1.7. These results confirm that conjugated RIBV is released from the carrier inside hepatic cells in a pharmacologically active form. Moreover, they indicate that the conjugated RIBV can display a stronger antiviral activity in the liver than the free dmg. To confirm the pharmacological equivalence in liver of doses of free and conjugated RIBV with comparable antiviral activity (Fig. 5), we measured the hepatic ['HIRIBVTP concentrations in mice administered with free or conjugated drug given by two daily im injections at the dosages of 30 and 13.2 pglg, respectively. Twenty-four hours after the first injection (and six hours after the second) in mice treated with 13.2

I

( 2

-

titer = 5.55 0.101 xdose

4 0

1 I

5

10

15

20

dose (pglglday) Fig. 5. Relationchip Irtwwn thc dally doses of frrc (0, 2nd jugarcd RlBV A ):ind M l l V titre (mean value I SL), 11. I l \ . e r

con.

pg/g/day of conjugated ["IRIBV and in those administered with 30 pglglday of free ['HIRIBV, the nmol of [3H]RIBVTP per g liver were 10.3 and 10.5, respectively. Discussion The experiments reported in this paper show that in mice treated with L-~oI~(LYS)-[~H]RIBV, the ratios between the levels of radioactivity in liver and those in RBC were 2 or 5 times higher than in animals which received the same dose of free [-IH]RIBVgiven im or po, respectively.The radioactivity in liver of mice treated with the conjugate was associated with pharmacologically active RIBV metabolites, as demonstrated by

425

G. Di Slsfano el al.

the experiments in which the hepatic concentrations of [3H]RIBVTP were measured and by those on the antiviral activity against MHV. Measurement of [3H]RIBVTP also in mouse RBC would have been important, since accumulation of RIBV in primate erythrocytes and the consequent haemolysis is considered to be caused by this metabolite which is only slowly degraded in human RBC ". However, the amounts of RIBVTP in mouse RBC are much smaller than in liver and we could not measure them. On the other hand, it can be reasoned that the ratio RIBVTP/total radioactivity in RBC should be higher in mice treated with the free drug than in those injected with the conjugate. The conjugate does not enter into erythrocytes and in mice injected with conjugated [3H]RIBV, the radioactivity detected in RBC is due to the drug and its metabolites which leave the hepatocytes after release from the canier within these cells 14. It has been shown 32 that in animals treated with RIBV the main compound exiting the cells is the de-ribosylated drug, which cannot be transformed into RIBVTP. The mechanism by which RIBV enhances the therapeutic activity of IFN in HCV infection is unknown; it has been hypothesized that the drug exerts this effect by acting on the immune system. If this is the case, a liver targeting of RIBV would not improve the efficacy of the drug. However, RTBV was found to synergistically increase the IFN activity against viruses growing in cells cultured in vitro, on which an action of the immune system can be excluded 33 34. This finding suggests that also in chronic hepatitis C the enhancing effect of RIBV is exerted inside the infected cells. Drawbacks in the clinical use of the conjugate might be the need to administer by the route and the replication of HCV also in blood mononuclear cells, in which the conjugate does not enter. As far as parenteral administrationis concerned, the increase in coniugate solubility obtained by adding ribose (which can-be safely given at high doses) 35 % an important achievement. In chronic type C hepatitis, RIBV is administered orally at the usual dose of 800-1200 mglday 2-6. The AUC (0-24 h) of hepatic [3H]RIBVTP was found to be 2-fold higher in mice injected with conjugated RIBV than in animals treated po with the same dose of the free drug (Fig. 4). This suggests that a similar pharmacological effect in liver could be ohtained by administering a smaller dose of conjugated RIBV, lower than that required by oral administration. Therefore, considering that the solubility of conjugated RIBV is 150 mglml it can be tentatively assumed that the pharmacologically active dose of the conjugate might he contained in a volume compatible with an im administration. Regarding HCV replication in blood mononuclear cells 36, it must be borne in mind

'

that the bulk of virus grows in hepatocytes 36-3s. Inhibition of its replication in these liver cells, as might occur with the use of the conjugate, should result in a drastic decrease in the virus production and in the number of newly infected cells. As a consequence the elimination of infected cells by the immune system should be facilitated. In conclusion, assuming that the results in mice could be applied to patients with chronic hepatitis C, the data reported here suggest that conjugated RIBV might display an antiviral activity against HCV in liver as the free drug, but with reduced haemolysis. One of the consequences of this side effect is an increase in iron deposition, particularly in hepatocytes "'. Since even modest amounts of excess iron in parenchymal liver cells could worsen viral hepatitis and decrease the likelihood of response to interferon 39, a reduced haemolysis might increase the efficacy of RIBV associated with IFN.

References

' Sidwell RW, H

u h JH, Khare GP, Allen LB, Witkowski JT, Robins RK. Broad-spectrum antiviral activity of virazole: l-B-D-ribofuranosyl-1,2,4-ti~ole-3-carboxamide. Science 1972:177:705-7. Brillanti S, Garson J, Foli M, Whitby K, Deaville R. Masci C. et al. A pilot study of combination therapy with ribavirin plus interferon alfa for interferon alfa-resistant chronic hepatitis C. Gastroenterology 1994:107:812-7. T h e m e l l o L, Cavalleno L, Bernardinello E, Guido M, Pontisso P, Albelti A. The effect of interferon alfa and ribavirin combination therapy in naive patients with chronic hepatitis C. J Hepatol 1995:23(Suppl2):8-12. Main J. Future studies of combination therapy f o chmnic ~ hepatioptimizing . .response .. . rates .. .for each hebititis C populatibn. J tis .. C . . Hepatol I ~ 5 : ~ 3 ( s u pZ):LIZ-b. p1 Lai MY, Kao JH, Yang PM, Wang IT, Chen PI, Chan KW, et al. Long term efficacy of ribavirin plus interferon alfa in the treatment of 1996:111:1107-12 ..chronic henatitir C -.. Gartroentemlnmr . ...,.. .. Schalm SW, Hansen BE, Chemello L, Bellobuono A, Brouwer JT,Weiland 0,et al. Ribavirin enhances the efficacy but not the adverse effects of interferon in chronic hepatitis C. eta-analysis of individual patients from European centers. J Hepatol 1997;26:961-6. 'Catlin DH, Smith RA, Samuels AI. 'T-ribavirin: distribution and pharmacokinetic studies in rats, baboons and man. In: Smith RA, Kirkpatrick W, editors. Ribavirin a broad spectrum antiviral agent. New York: Academic Press, 1980:83-98. Laskin OL, Longstreth JA, Hart CC, Scavuzzo D, Kalman CM, Connor ID, et al. Ribavirin disposition in high-risk patients for acquired immunodeficiency syndrome. Clin Pharmacol Ther 1987;41:546-55. r). R!,rcgljr :\\l. I l x o n FIR. Klciner DE, Hoofnagle 11. Incruw In ~ L . P : I I I L 11011\lore\ iollo~ving prolonged therap) with nhavinn 11, p.8!lcnt\ nlrl,