sign in sign up
<<
Home , Antiviral drug

ANTIMICROBIAL AGENTS AND , Feb. 1996, p. 514–519 Vol. 40, No. 2 0066-4804/96/$04.00ϩ0 Copyright ᭧ 1996, American Society for

In Vitro Potency of Inhibition by Antiviral of Hematopoietic Progenitor Colony Formation Correlates with Exposure at Hemotoxic Levels in Human Immunodeficiency -Positive Humans

RONNA E. DORNSIFE* AND DEVRON R. AVERETT Division of Experimental , Burroughs Wellcome Co., Research Triangle Park, North Carolina

Received 9 June 1995/Returned for modification 1 September 1995/Accepted 16 November 1995

Inhibition of in vitro colony formation of human hematopoietic progenitors (CFU-granulocyte-macrophage, burst-forming unit-erythroid) by the antiviral drugs , , , , , , , and acyclovir was measured. Significant correlations between in vitro 50% inhibitory concentrations and the daily human exposures (area under the concentration-time curve from 0 to 24 h; in micromolar ⅐ hour) of these chronically administered drugs in human immunodeficiency virus-positive patients that induced neutropenia or anemia were demonstrated by both linear regression and Spearman rank-order analyses. These quantitative correlations allow estimation of the exposure at which marrow toxicity may occur with candidate compounds.

The results of in vitro assays are used to understand and numbers in phosphate-buffered saline-treated controls from predict in vivo observations. Candidate antiviral compounds 142 experiments with samples from 61 different donors were 33 are typically evaluated in vitro for both antiviral activity and Ϯ 14 (standard deviation) per well for CFU-GM and 77 Ϯ 29 cellular toxicity to identify selective compounds (33). Bone per well for BFU-E (from experiments cited for AZT; Table marrow toxicity is the principal dose-limiting toxicity associ- 1). ated with a number of antiviral drugs (30, 49). Thus, candidate Antiviral drugs were obtained through the Wellcome com- antiviral compounds have been tested in colony-forming assays pound registry. Inhibition of colony growth from bone marrow for human marrow progenitors as an in vitro model for hema- progenitors by each compound was determined over an appro- tologic toxicity in humans (32, 38, 54). priate range of concentrations. Each was evaluated in a Colony-forming assays for bone marrow progenitors are use- minimum of four experiments with a minimum of four differ- ful for measuring in vivo hematopoietic responses to both ent marrow donors (Table 1). The percentage of control values disease and chemotherapeutic agents and serve as key in vitro was calculated by dividing the average of the colony counts model systems in experimental hematology (4, 6, 36). Parallels from drug-exposed samples (n ϭ 2 wells) by the average colony between in vitro and in vivo bone marrow toxicity have been counts from control samples (n ϭ 6 wells, saline solvent) in documented for individual chemotherapeutic agents, including each independent experiment. The percentage of control val- antiviral drugs (12, 29, 43, 54). Quantitative relationships be- ues for each drug at each concentration was averaged across tween in vitro progenitor toxicity and in vivo plasma drug levels independent experiments. Dose-response curves were derived associated with decreased numbers of peripheral cells from these data. with certain therapeutic agents have also been postulated (28, Calculation of IC s of antiviral drugs. The 50% inhibitory 32, 37, 44, 56, 58). However, no study which has demonstrated 50 concentrations (IC s) were calculated from the dose-response a statistical correlation between in vitro and in vivo bone mar- 50 curves for each drug. IC s were determined by either the log row toxicity with a survey of drugs has been reported. In this 50 drug concentration-versus-percent inhibition algorithm or the report, we present a comprehensive analysis of quantitative correlations between in vitro and in vivo hematopoietic toxicity Hill equation (15, 40). The error associated with the estimation for clinically significant antiviral drugs. of the IC50sisϮthe standard error (i.e., the fit of data to the In vitro bone marrow progenitor colony-forming assay. The line; Table 1). Estimates of the IC50 for a particular drug in vitro toxicities of antiviral drugs to human bone marrow varied less than twofold when the drugs were tested in separate progenitors were assessed in a standardized colony-forming replicate experiment sets (i.e., interassay variation). assay by previously published methods (15). Zidovudine Pharmacokinetic data. Daily drug exposure (area under the (AZT) samples at 10, 1, and 0.1 ␮M were included as stan- concentration-time curve [AUC] from 0 to 24 h [AUC0–24]) dards in every experiment. CFU-granulocyte-macrophage data were derived from clinical pharmacokinetic assessments (CFU-GM) and burst-forming unit-erythroid (BFU-E) colo- in human immunodeficiency virus (HIV)-infected patients (5, nies were counted at ϫ40 and ϫ100 magnifications, respec- 13, 20, 24, 31, 34, 35, 48). The methods used by the original tively, on a phase-contrast, inverted microscope on days 11 to investigators to calculate AUC varied, but most included ac- 13 of the assay (the setup day was day 0). Average colony cepted applications of trapezoidal methods. AUC0–24 values were calculated from available pharmacokinetic data to match as closely as possible the dosage regimen used in trials that * Corresponding author. Mailing address: Glaxo Wellcome Inc., reported hematologic toxicity (see legend to Fig. 1). When Rm. 715, Research Commons II Building, 3030 Cornwallis Rd., Re- available, daily exposure was estimated from the sustained expo- search Triangle Park, NC 27709. Phone: (919) 483-9471. Fax: (919) sure at steady state from multiple dosings. If only single-dose 315-5243. Electronic mail address: ronna [email protected]. exposure data were available, the AUC0–24 was estimated by

514 VOL. 40, 1996 NOTES 515

TABLE 1. In vitro toxicities of antiviral drugs to human bone 100 marrow progenitors Percent observed ϭ 50% effect n a 1 ϩ IC50 (␮M) No. of No. of ͩoriginal AUC0–24ͪ Drug Reference(s) CFU-GM BFU-E expt donors Thus, if n equal to infinity equals the observed incidence, then

Alovudine 0.60 Ϯ 0.3 0.07 Ϯ 0.01 4 4 11 n equal to 1 (Hill plot) describes the largest possible AUC0–24 ddC 1.6 Ϯ 0.3 0.70 Ϯ 0.1 11 10 60 value equal to a 50% effect or incidence. Original AUC0–24 AZT 7 Ϯ 6 0.4 Ϯ 0.1 142 61 15, 59 data were therefore adjusted by the following formula (as Ganciclovir 15 Ϯ 339Ϯ10 18 15 7, 59 above, n is equal to 1 and is restated) to approximate a 50% Stavudine 70 Ϯ 10 30 Ϯ 10 5 4 Didanosine Ͼ400b 34 Ϯ 48615 incidence of the neutropenia or anemia. Lamivudine 250 Ϯ 8 180 Ϯ 26625 Acyclovir ACV 220 Ϯ 50 280 Ϯ 50 15 13 59 Adjusted AUC0–24 100% Ϫ percent incidence of cytopenia a Values are means Ϯ standard errors of the means. ϭ b Depicted and analyzed as equal to 400 ␮M in Fig. 1A and B and Table 2. ͩ percent incidence of cytopenia ͪ

⅐ original AUC0–24 multiplying the AUC from time zero to infinity (AUC0–ϱ) for The adjusted data provided a theoretical exposure equal to a a single dose by the number of doses that the patients received 50% effect, thus equalizing the in vivo data among all the drugs per day. Estimates were made from data in which the dose and permitting analysis with in vitro IC50s. linearity and proportionality of each drug for the dose calcu- For comparative purposes, the data are also presented for lated were confirmed. Estimates were also confirmed by com- the anti- virus drug acyclovir, which at elevated parison with available pharmacokinetic data for non-HIV-in- doses does not induce hematologic toxicity (64). No clinically fected patients with normal renal and liver functions for each significant changes in hematologic parameters were reported drug (data not shown). The standard of 70 kg for human body in the previous study (64), and the patients on acyclovir had a mass was used for conversion purposes in our pharmacokinetic survival benefit. In another study with the same dose of acy- calculations. For our analysis, we assumed that the ratio of the clovir used in combination with AZT (10), data averaged from drug concentration in bone marrow to that in plasma was the patients with AIDS and AIDS-related complex resulted in less same for all drugs. The AUC was expressed as micromolar ⅐ than background level increases in hematologic toxicity over hour. the percentages observed in patients treated with AZT alone Hematologic toxicity of antiviral drugs in HIV-positive hu- (Ͻ10% increase in neutropenia; Ͻ5% increase in anemia). In mans from clinical studies. Hematologic toxicity has been doc- addition, the pharmacokinetic data available for acyclovir in umented in clinical trials with the antiviral drugs analyzed in HIV-infected patients cited here were from a study in which our study (1, 9, 24, 41, 46–48, 63). Assessment of hematologic acyclovir was administered in combination with AZT (13). toxicity caused by a particular antiviral drug, however, is com- Regression and rank-order analyses. Linear regression and plicated by the multiple factors that induce bone marrow sup- Spearman rank-order correlation analyses were performed by pression in HIV-infected patients (14, 16). The largest popu- using the STATVIEW II program (Abacus Concepts, Inc., lation allowing estimation of a background rate of hematologic Berkeley, Calif.) (Table 2). Statistical analyses were also per- dysfunction was a group of 107 placebo-dosed, HIV-infected formed on AUC0–24 data adjusted to approximate a 50% in- patients with advanced symptoms from early trials with AZT cidence of the corresponding level of cytopenia for each drug (47). On the basis of the data from this group of AIDS patients, at each exposure (see equation above). Because of the large 3 we used 10% neutropenia (defined as Ͻ1,000 cells per mm ) range of potencies of these compounds, all in vitro IC50s and and 5% anemia (defined as Ͻ7.5 g of hemoglobin per dl or AUC0–24 data were log transformed prior to statistical corre- transfusions) as background levels of hematologic suppression. lation analyses. Only percentages greater than these were considered attribut- Six antiviral drugs inducing severe neutropenia in chroni- able to drug. cally dosed HIV-positive patients were analyzed: alovudine, Because the present study is in part a retrospective analysis zalcitabine (ddC), stavudine, and lamivudine at elevated doses; of published clinical trials, a number of parameters were not AZT at elevated and current doses; and ganciclovir at cur- controllable. For example, both the HIV disease stage and the rently recommended doses. All six drugs induce neutropenia in absolute number of patients evaluated in a particular study 13 to 60% of patients, an incidence up to sixfold above the differed; thus, the relative percentage of patients affected with background incidence. The in vitro IC50s for inhibition of hematologic toxicity varied proportionally. For this reason, we CFU-GM bone marrow progenitors were compared with the matched clinical parameters as closely as possible in our anal- AUC0–24 values at which neutropenias were observed (Fig. 1A; ysis, including the criteria used to define anemia or neutrope- Table 1). The percentage of patients affected with neutropenia nia. upon chronic dosing with each drug at the associated level of The percentages of patients affected with neutropenia or exposure is included parenthetically in Fig. 1A. Regression anemia varied among the different drugs and exposures; for analyses were performed for original and adjusted data (Fig. this reason original AUC0–24 data were analyzed and com- 1A; Table 2). The regression analyses described highly signif- pared with AUC0–24 data adjusted to approximate a 50% in- icant positive linear correlations between in vitro toxicity to cidence of the corresponding level of cytopenia for each drug CFU-GM progenitors and the drug exposure inducing neutro- at each exposure. A logistic equation based on biochemical penia in patients dosed chronically with these drugs (Fig. 1A; binding curve kinetics was used to determine a theoretical 50% Table 2). In addition, Spearman rank-order analysis of these incidence of cytopenia at a particular level of exposure (40, 61). same data confirmed the statistical significance of this corre-

The adjusted AUC0–24 values are based on the assumption that lation (Table 2). an increase or a decrease in drug exposure is proportional to We performed similar analyses for four antiviral drugs which the increase or decrease in toxicity. induce anemia when dosed chronically (Fig. 1B): alovudine, 516 NOTES ANTIMICROB.AGENTS CHEMOTHER.

FIG. 1. (A) Correlation of in vitro CFU-GM toxicity with in vivo neutropenia. (B) Correlation of in vitro bone marrow BFU-E toxicity with in vivo anemia. The estimated AUC0–24 values were as follows: alovudine (FLT), 1.64, 2.46, and 3.42 ␮M ⅐ h; ddC, 6.36 ␮M ⅐ h; AZT, 12.2, 32.3, and 40.4 ␮M ⅐ h; ganciclovir (GCV), 120 ␮M ⅐ h; stavudine (d4T), 88.7 ␮M ⅐ h; didanosine (ddI), 29.9 ␮M ⅐ h; lamivudine (3TC), 127 ␮M ⅐ h; and acyclovir (ACV), 62.4 ␮M ⅐ h. stavudine, and didanosine at elevated doses and AZT at both ground of 5% anemia observed in placebo-dosed AIDS pa- elevated doses and the currently recommended dose (2, 3, 21, tients. The regression analyses for the erythroid data, both with 23, 47, 49). These drugs induce moderate to severe anemia in original data and with adjusted data, yielded linear correlations 15 to 75% of patients tested, which is 3- to 15-fold the back- which were statistically significant (Fig. 1B; Table 2). Spear- VOL. 40, 1996 NOTES 517

a TABLE 2. Correlation analyses of in vitro to in vivo hematologic toxicity and in vitro IC50s from independent laboratories

Linear regression analysis Spearman Comparison rank-order rr2 P P value

CFU-GM IC50 versus neutropenia (Fig. 1A) Data (n ϭ 9; six different drugs) 0.913 0.834 0.0006b 0.007b Adjusted data 0.938 0.880 0.0002b 0.007b

BFU-E IC50 versus anemia (Fig. 1B) Data (n ϭ 8; four different drugs) 0.775 0.601 0.02c 0.04c Adjusted data 0.833 0.694 0.01b 0.01b

Dornsife et al. (see Table 1) versus Sommadossi et al. (27, 50, 54–56)

CFU-GM IC50s Drugs that induced neutropenia (n ϭ 6) 0.930 0.866 0.007b 0.04c All drugs reported (n ϭ 8) 0.952 0.906 0.001b 0.02c

Dornsife et al. (see Table 1) versus Sommadossi et al. (27, 50, 54–56)

BFU-E IC50s Drugs that induced anemia (n ϭ 4) 0.920 0.846 0.08 0.08 All drugs reported (n ϭ 7, except acyclovir) 0.845 0.714 0.02c 0.05c

a r is the correlation coefficient. P is the associated probability for correlation if the null hypothesis (no correlation between the two variables) is rejected at the indicated significance level (P Յ 0.05 or 0.01). Confidence intervals for the mean and the slope of a regression line allow estimation of the likelihood that a true value would lie in the selected range on the basis of the original samples and a chosen level of probability (e.g., 90%; see insets in Fig. 1A and B). Sets of quantitative data are expressed in terms of rank order for Spearman correlation analyses (57). For the Spearman rank-order analysis, P is a two-tailed probability. b Significant at the level of P Յ 0.01. c Significant at the level of P Յ 0.05. man rank-order analyses of these same data confirmed this cation of 90% confidence intervals to the mean and the slope result (Table 2). of the regression (original AUC0–24 values) suggests that ddC To corroborate these findings on toxicity, we compared our is an exception to the model presented (inset to Fig. 1B). data (Table 1) with the results of a similar assay used by an Although both our data and those of other investigators indi- independent laboratory and analyzed the results for statistical cate that ddC is a potent inhibitor of BFU-E, no anemias were correlations. Representative in vitro IC50s of each of the anti- reported in the patients (n ϭ 5) given ddC at the elevated viral drugs published by Sommadossi and coworkers were in- dosage analyzed (0.54 mg/kg/day Ϸ 38 mg/day) (19, 26, 27, 32, cluded in the analyses (27, 50, 54, 55, 56). Statistically signifi- 63). Because of the dose-limiting peripheral neuropathy asso- cant linear and rank-order correlations were found for the ciated with ddC treatment, considerably lower doses (2.25 mg/ CFU-GM data from the two laboratories (Table 2). A corre- day) are currently used in combination therapy with AZT. This lation for the BFU-E data between the two laboratories was dose of ddC, as a monotherapy, is associated with anemia in not statistically significant for the drugs that induced anemia, approximately 5% of patients (51). This level of anemia cannot but it was significant for a linear correlation when the available be distinguished from background incidence levels of anemia data for all drugs were included. In view of these correlations, in AIDS patients. Therefore, the regression analysis with the it was not surprising that analysis for correlations of the in vitro adjusted data may more accurately describe the linear corre- data of Sommadossi and colleagues with drug exposure and in lation for the incidence of severe anemia (Fig. 1B; Table 2). vivo toxicity findings provided results that closely paralleled Comparison of our in vitro IC50s with data from an inde- those presented above for CFU-GM and neutropenia, whereas pendent laboratory demonstrated both a significant linear cor- the pattern for BFU-E and anemia was generally similar (data relation and strong to significant rank-order correlations for not shown). both progenitors. These results highlight the fact that, despite

Relationships between in vitro and in vivo toxicity to bone differences in the absolute IC50s of compounds from different marrow have been described for anticancer and antiviral drugs laboratories, the data generated in such assays can be highly (12, 18, 43, 58). Sommadossi and colleagues were among those correlative. Thus, the data from either laboratory can be used who initially applied in vitro colony-forming assays to assess to derive similar conclusions on relative toxicity, as was pro- the toxicity of antiviral drugs (38, 45, 54, 62). Sommadossi and posed previously (55). Carlisle were the first to publish data on AZT (54). These and Data on adverse events, including hematologic toxicities, other investigators have suggested relationships between in from clinical trials are most often reported as the number of vitro potency, peak or sustained concentrations in plasma, and percentage of affected patients in each dose group and the observed levels of cytopenia in humans with individual agents severity of the toxicity. Our analyses with calculated drug ex- (22, 52–56, 65). We present here a rigorous quantitative anal- posures and different incidences of cytopenias were also com- ysis of these hypotheses and test for a statistical correlation pared with analyses based on theoretical drug exposures for between in vitro and in vivo bone marrow toxicity data with a each drug estimated to be equal to a 50% incidence of patients survey of drugs which has not been previously reported. We affected by either neutropenia or anemia (adjusted data). This found relationships between in vitro and in vivo hematologic approach was taken in an effort to standardize the percentage toxicities that were statistically significant. of patients receiving the different drugs affected and allow for The absence of anemia (Ͼ5%) reported for both ddC at the analysis of correlation with a 50% effect in vitro. Statistically elevated dose examined and ganciclovir at the recommended significant linear and rank-order correlations were observed dose suggest that exceptions may exist for this model. Appli- with both the calculated drug exposures and the adjusted data 518 NOTES ANTIMICROB.AGENTS CHEMOTHER. for drug exposure. However, if the data were available, drug 11. Daluge, S. M., D. J. M. Purifoy, P. M. Savina, M. H. St. Clair, N. R. Parry, exposure measurements from those patients who experienced I. K. Dev, P. Novak, K. M. Ayers, J. E. Reardon, G. B. Roberts, J. A. Fyfe, M. R. Blum, D. R. Averett, R. E. Dornsife, B. A. Domin, R. Ferone, D. A. a 50% decrease in neutrophil numbers or hemoglobin from the Lewis, and T. A. Krenitsky. 1994. 5-Chloro-2Ј,3Ј-dideoxy-3Ј-fluorouridine baseline might provide a more relevant comparison with the in (935U83), a selective anti-human immunodeficiency virus agent with an vitro IC50 toxicity assessment (48, 58). improved metabolic and toxicological profile. Antimicrob. Agents Che- The antiviral drugs tested in our study manifest different mother. 38:1590–1603. 12. Deldar, A., and C. E. Stevens. 1993. Development and application of in vitro types of dose-limiting toxicities in humans. In vitro toxicity models of hematopoiesis to drug development. Toxicol. Pathol. 21:231–240. assessment of a candidate antiviral compound with primary 13. de Miranda, P., S. Weller, M. Maha, A. R. Lifson, M. A. Piccini, G. W. human marrow progenitors cannot predict if marrow toxicity in Rutherford, H. Hollander, and M. R. Blum. 1988. of humans will be the dose-limiting toxicity. However, predictions zidovudine (ZDV) and acyclovir (ACV) during concurrent drug administra- tion in asymptomatic men with HIV . abstr. 3136, p. 253. In Program of sustained drug exposures at which a candidate compound and abstracts, vol. 1. IV International Conference on AIDS. may induce neutropenia or anemia in humans can be made. 14. Donahue, R. E., M. M. Johnson, L. I. Zon, S. C. Clark, and J. E. Groopman. The correlation between in vitro and in vivo marrow toxicity 1987. Suppression of in vitro haematopoiesis following human immunodefi- observed with these antiviral agents was striking. This finding ciency virus infection. Nature (London) 326:200–203. 15. Dornsife, R. E., M. H. St. Clair, A. T. Huang, T. J. Panella, G. W. Koszalka, was especially notable because of the numerous variables in- C. L. Burns, and D. R. Averett. 1991. Anti-human immunodeficiency virus volved in in vitro and in vivo hematologic toxicity assessment, synergism by zidovudine (3Ј-azidothymidine) and didanosine (dideoxyi- the multifactorial nature of bone marrow suppression in HIV- nosine) contrasts with their additive inhibition of normal human marrow infected patients, and the possibly, different mechanisms by progenitor cells. Antimicrob. Agents Chemother. 35:322–328. 16. Doweiko, J. P., and J. E. Groopman. 1994. Hematological consequences of which these drugs may exert their toxic effects. Thus, the HIV infection, p. 617–628. In S. Broder, T. C. Merigan, Jr., and D. Bolognesi present analysis confirms the roles of these in vitro assays in (ed.), Textbook of AIDS medicine. The Williams & Wilkins Co., Baltimore. assessments of the toxicities of antiviral compounds during 17. Drach, J. C., L. B. Townsend, M. R. Nassiri, S. R. Turk, L. A. Coleman, R. V. preclinical evaluation in selecting compounds with low poten- Devivar, G. Genzlinger, E. D. Kreske, T. E. Renau, A. C. Westerman, C. Shipman, Jr., K. K. Biron, R. Dornsife, and E. R. Kern. 1992. Benzimidazole tial for hematologic toxicity (8, 11, 17, 25, 42). ribonucleosides: a new class of antivirals with potent and selective activity against human . Antiviral Res. 17(Suppl. 1):49 (abstr. 12). We acknowledge Carlos de Castro, Tim Panella, and Andrew T. 18. Drusano, G. L., F. M. Balis, S. R. Gitterman, and P. A. Pizzo. 1994. Quan- Huang of Duke University Medical Center for collection of bone titative relationships between zidovudine exposure and efficacy and toxicity. marrow samples; Debra Sanders, Gregg Workman, Amy Tanner, and Antimicrob. Agents Chemother. 38:1726–1731. Joseph Paisley for technical support; David Porter for statistical ad- 19. Du, D. L., D. A. Volpe, C. K. Grieshaber, and M. J. Murphy, Jr. 1990. In vitro vice, and Steve Good, Richard Miller, and Tom Zimmerman for crit- myelotoxicity of 2Ј,3Ј-dideoxynucleosides on human hematopoietic progen- itor cells. Exp. Hematol. 18:832–836. ical discussions. 20. Dudley, M. N., K. K. Graham, S. Kaul, S. Geletko, L. Dunkle, M. Browne, This work was supported by Burroughs Wellcome Co. and K. Mayer. 1992. Pharmacokinetics of stavudine in patients with AIDS or AIDS-related complex. J. Infect. Dis. 166:480–485. REFERENCES 21. Dunkle, L., A. Cross, R. Gugliotti, R. Martin, M. Browne, and H. Murray. 1. Allan, J. D., K. J. Connolly, H. Fitch, L. Jackson-Pope, C. McLaren, R. 1990. Dose-escalating study of safety and efficacy of dideoxy-didehydrothy- Canetta, and J. E. Groopman. 1993. Long-term follow-up of didanosine midine (d4T) for HIV infection. Antiviral Res. April(Suppl. 1):116 (abstr. administered orally twice daily to patients with advanced human immuno- 145). deficiency virus infection and hematologic intolerance of zidovudine. Clin. 22. Faraj, A., D. A. Fowler, E. G. Bridges, and J.-P. Sommadossi. 1994. Effects Infect. Dis. 16(Suppl. 1):S46–S51. of 2Ј,3Ј-dideoxynucleosides on proliferation and differentiation of human 2. Barditch, P., B. Polsky, C. van der Horst, F. Ducanson, O. Kuye, D. Ganes, pluripotent progenitors in liquid culture and their effects on mitochondrial R. Faulkner, R. Raasch, D. Armstrong, B. Petty, B. Rae, A. Yacobi, C. DNA synthesis. Antimicrob. Agents Chemother. 38:924–930. Flexner, and R. Desjardins. 1992. Hematologic toxicity of FLT appears to be 23. Fischl, M. A., C. B. Parker, C. Pettinelli, M. Wulfsohn, M. S. Hirsch, A. C. AUC dependent. Clin. Pharm. Ther. 51:170 abstr. Oll-1). Collier, D. Antoniskis, M. Ho, D. D. Richman, E. Fuchs, T. C. Merigan, R. C. 3. Barditch-Crovo, P., D. Ganes, F. Duncanson, C. van der Horst, B. Polsky, B. Reichman, J. Gold, N. Steigbigel, G. S. Leoung, S. Rasheed, A. Tsiatis, and Petty, R. Faulkner, O. Kuye, A. Yacobi, P. Lietman, C. Norden, and C. the AIDS Clinical Trials Group. 1990. A randomized controlled trial of a Flexner. 1992. Early markers of hematologic toxicity with FLT therapy, abstr. reduced daily dose of zidovudine in patients with the acquired immunode- WeB 1012, p. We 47. In Abstract book, vol. 1. VIII International Conference ficiency syndrome. N. Engl. J. Med. 323:1009–1014. on AIDS/III STD World Congress. Harvard AIDS Institute, Cambridge. 24. Flexner, C., C. van der Horst, M. A. Jacobson, W. Powderly, F. Duncanson, 4. Beris, P., and P. A. Miescher. 1988. Hematological complications of antiin- D. Ganes, P. A. Barditch-Crovo, B. G. Petty, P. A. Baron, D. Armstrong, P. fectious agents. Semin. Hematol. 25:123–139. Bricmont, O. Kuye, A. Yacobi, R. DesJardins, and B. Polsky. 1994. Rela- 5. Blum, M. R., S. H. T. Liao, S. S. Good, and P. de Miranda. 1988. Pharma- tionship between plasma concentrations of 3Ј-deoxy-3Ј-fluorothymidine cokinetics and bioavailability of zidovudine in humans. Am. J. Med. (alovudine) and antiretroviral activity in two concentration-controlled trials. 85(Suppl. 2A):189–194. J. Infect. Dis. 170:1394–1403. 6. Bradley, T. R., and D. Metcalf. 1966. The growth of mouse bone marrow 25. Furman, P. A., M. Davis, D. C. Liotta, M. Paff, L. W. Frick, D. J. Nelson, cells in vitro. Aust. J. Exp. Biol. Med. Sci. 44:287–299. R. E. Dornsife, J. A. Wurster, L. J. Wilson, J. A. Fyfe, J. V. Tuttle, W. H. 7. Chamberlain, S. D., K. K. Biron, R. E. Dornsife, D. R. Averett, L. Miller, L. Condreay, D. R. Averett, R. F. Schinazi, and G. R. Painter. 1992. Beauchamp, and G. W. Koszalka. 1994. An enantiospecific synthesis of the The anti- activities, cytotoxicities, and anabolic profiles of the (Ϫ) human cytomegalovirus antiviral agent [(R)-3-((2-amino-1,6-dihydro-6-oxo- and (ϩ) enantiomers of cis-5-fluoro-1-[2-(hydroxymethyl)-1,3-oxthiolane-5- 9H-purin-9-yl)methoxy)-4-hydroxybutyl]phosphonic acid. J. Med. Chem. 37: yl]. Antimicrob. Agents Chemother. 36:2686–2692. 1371–1377. 26. Ganser, A., J. Greher, B. Volkers, S. Staszewski, and D. Hoelzer. 1989. 8. Ching, S. V., K. M. Ayers, R. E. Dornsife, G. L. Grebe, and J. L. Howard. Inhibitory effect of azidothymidine, 2Ј-3Ј-dideoxyadenosine, and 2Ј-3Ј 1994. Nonclinical toxicology and in vitro toxicity studies with the novel dideoxycytidine on in vitro growth of hematopoietic progenitor cells from anti-HIV agent (1S,4R)-4-[2-amino-6-(cylcopropylamino)-9H-purin-9-yl]-2- normal persons and patients with AIDS. Exp. Hematol. 17:321–325. cyclopentene-1-methanol (1592U89) succinate, abstr. I88, p. 92. In Program 27. Gosselin, G., R. F. Schinazi, J.-P. Sommadossi, C. Mathe, M.-C. Bergogne, and abstracts of the 34th Interscience Conference on Agents A.-M. Aubertin, A. Kirn, and J.-L. Imbach. 1994. Anti-human immunodefi- and Chemotherapy. American Society for Microbiology, Washington, D.C. ciency virus activities of the ␤-L-enantiomer of 2Ј,3Ј-dideoxycytidine and its 9. Connolly, K. J., J. D. Allan, H. Fitch, L. Jackson-Pope, C. McLaren, R. 5-fluoro derivative in vitro. Antimicrob. Agents Chemother. 38:1292–1297. Canetta, and J. E. Groopman. 1991. Phase I study of 2Ј-3Ј-dideoxyinosine 28. Henderson, J. R. 1989. Use of ganciclovir in the treatment of cytomegalo- administered orally twice daily to patients with AIDS or AIDS-related com- virus . Br. J. Clin. Pract. 43:231–237. plex and hematologic intolerance to zidovudine. Am. J. Med. 91:471–478. 29. Horikoshi, A., and M. J. Murphy, Jr. 1982. Comparative effects of chemo- 10. Cooper, D. A., P. O. Pehrson, C. Pedersen, M. Moroni, E. Oksenhendler, W. therapeutic drugs on human and murine hematopoietic progenitors in vitro. Rozenbaum, N. Clumeck, V. Faber, W. Stille, B. Hirschel, C. Farthing, R. Chemotherapy (Basel) 28:480–501. Doherty, J. M. Yeo, and a European-Australian Collaborative Group. 1993. 30. Irons, R. D. (ed.). 1985. Toxicology of the blood and bone marrow. Target The efficacy and safety of zidovudine alone or as cotherapy with acyclovir for organ toxicology series. Raven Press, New York. the treatment of patients with AIDS and AIDS-related complex: a double- 31. Jacobson, M. A., P. deMiranda, D. M. Cederberg, T. Burnette, E. Cobb, blind, randomized trial. AIDS 7:197–207. H. R. Brodie, and J. Mills. 1987. Human pharmacokinetics and tolerance of VOL. 40, 1996 NOTES 519

oral ganciclovir. Antimicrob. Agents Chemother. 31:1251–1254. rative Working Group. 1987. The toxicity of azidothymidine (AZT) in the 32. Johnson, M., T. Caiazzo, J.-M. Molina, R. Donohue, and J. Groopman. 1988. treatment of patients with AIDS and AIDS-related complex: a double-blind Inhibition of bone marrow myelopoiesis and erythropoiesis in vitro by anti- placebo-controlled trial. N. Engl. J. Med. 317:192–197. retroviral nucleoside derivatives. Br. J. Haemtol. 70:137–141. 50. Schinazi, R. F., A. McMillan, D. Cannon, R. Mathis, R. M. Lloyd, A. Peck, 33. Kern, E. R. 1990. Preclinical evaluation of antiviral agents: in vitro and J.-P. Sommadossi, M. St. Clair, J. Wilson, P. A. Furman, G. Painter, W.-B. animal model testing, p. 87–94. In G. J. Galasso, R. J. Whitley, and T. C. Choi, and D. C. Liotta. 1992. Selective inhibition of human immunodefi- Merigan (ed.), Antiviral agents and viral diseases of man, 3rd ed. Raven ciency by racemates and enantiomers of cis-5-fluoro-1-[2-(hydroxy- Press, New York. methyl)-1,3-oxathiolan-5-yl]cytosine. Antimicrob. Agents Chemother. 36:2423– 34. Klecker, R. W., Jr., J. M. Collins, R. C. Yarchoan, R. Thomas, N. McAtee, S. 2431. Broder, and C. E. Myers. 1988. Pharmacokinetics of 2Ј,3Ј-dideoxycytidine in 51. Scott, R., S. Follansbee, R. Olson, R. Pollard, W. Reiter, M. Salgo, et al. patients with AIDS and related disorders. J. Clin. Pharmacol. 28:837–842. 1993. Safety and tolerance of zalcitabine (ddC, Hivid᭨) in a double-blind 35. Knupp, C. A., W. C. Shyu, R. Dolin, F. T. Valentine, C. McLaren, R. R. comparative trial (ACTG 114: N3300), abstr. PO-B26-2115, p. 487. In Pro- Martin, K. A. Pittman, and R. H. Barbhaiya. 1991. Pharmacokinetics of gram and abstracts, vol. 1 IX International Conference on AIDS. Wellcome didanosine in patients with acquired immunodeficiency syndrome or ac- Foundation Ltd., London. quired immunodeficiency syndrome-related complex. Clin. Pharmacol. Ther. 52. Sommadossi, J.-P. 1993. Nucleoside analogs: similarities and differences. 49:523–535. Clin. Infect. Dis. 16(Suppl. 1):S7–S15. 36. Marsh, J. C. 1976. The effects of cancer chemotherapeutic agents on normal 53. Sommadossi, J. P., D. W. Barnes, L. R. Miller, M. A. Markiewicz, and R. J. hematopoietic precursor cells: a review. Cancer Res. 36:1853–1882. Whitley. 1987. Novel pharmacologic strategies in the treatment of life-threat- 37. Matthews, T., and R. Boehme. 1988. Antiviral activity and mechanism of ening infections: clinical experience with 9-(1,3-dihydroxy-2-propoxy-meth- action of ganciclovir. Rev. Infect. Dis. 10(Suppl. 3):490–494. yl)- (DHPG). Clin. Res. 35:491A. 38. McGuffin, R. W., F. M. Shiota, and J. D. Meyers. 1980. Lack of toxicity of 54. Sommadossi, J.-P., and R. Carlisle. 1987. Toxicity of 3Ј-azido-3Ј-deoxythy- acyclovir to granulocyte progenitor cells in vitro. Antimicrob. Agents Che- midine and 9-(1,3-dihydroxy-2-propoxymethyl)guanine for normal human mother. 18:471–473. hematopoietic progenitor cells in vitro. Antimicrob. Agents Chemother. 39. Mitsuya, H., and R. Yarchoan. 1994. Development of antiretroviral therapy 31:452–454. for AIDS and related disorders, p. 721–742. In S. Broder, T. C. Merigan, Jr., 55. Sommadossi, J.-P., R. F. Schinazi, C. K. Chu, and M.-Y. Xie. 1992. Com- and D. Bolognesi (ed.), Textbook of AIDS medicine. The Williams & parison of cytotoxicity of the (Ϫ)- and (ϩ)-enantiomer of 2Ј,3Ј-dideoxy-3Ј- Wilkins Co., Baltimore. thiacytidine in normal human bone marrow progenitor cells. Biochem. Phar- 40. Monod, J., J.-P. Changeux, and F. Jacob. 1963. Allosteric and macol. 44:1921–1925. cellular control systems. J. Mol. Biol. 6:306–329. 56. Sommadossi, J.-P., Z. Zhu, R. Carlisle, M.-Y. Xie, and D. A. Weidner. 1990. 41. Murray, H. W., K. E. Squires, W. Weiss, S. Sledz, H. S. Sacks, J. Hassett, A. Pharmacologic studies of active against the human immunode- Cross, R. E. Anderson, and L. M. Dunkle. 1995. Stavudine in patients with ficiency virus. Ann. N. Y. Acad. Sci. 616:356–366. AIDS and AIDS-related complex: AIDS clinical trials group 089. J. Infect. 57. Spearman, C. 1904. The proof and measurement of association between two Dis. 171(Suppl. 2):S123–S130. things. Am. J. Psychcol. 15:72–101. 42. Painter, G. R., M. H. St. Clair, P. de Miranda, D. Reynolds, S. Ching, R. 58. Stretcher, B. N., A. J. Pesce, P. T. Frame, K. A. Greenberg, and D. S. Stern. Dornsife, D. J. Livingston, S. Pazhanisamy, and R. Tung. 1995. An overview 1994. Correlates of zidovudine phosphorylation with markers of HIV disease of the preclinical development of the HIV inhibitor VX-478 progression and drug toxicity. AIDS 8:763–769. (141W94), abstr. LB5, p. 167. In Program and abstracts of 2nd National 59. Szczech, G. 1996. Preclinical development of antiviral drugs. Clin. Infect. Conference on Human and Related Infections. American So- Dis. 22:355–360. ciety for Microbiology Books International, Herndon, Va. 60. van Draanen, N. A., M. Tisdale, N. R. Parry, R. Jansen, R. E. Dornsife, J. V. 43. Parchment, R. E., M. Huang, and C. L. Erickson-Miller. 1993. Roles for in Tuttle, D. R. Averett, and G. W. Koszalka. 1994. Influence of stereochemistry vitro myelotoxicity tests in preclinical drug development and on antiviral activities and resistance profiles of dideoxycytidine nucleosides. planning. Toxicol. Pathol. 21:241–250. Antimicrob. Agents Chemother. 38:868–871. 44. Parchment, R. E., D. A. Volpe, P. M. LoRusso, C. L. Erickson-Miller, M. J. 61. van Holde, K. E. (ed.). 1971. Multiple equilibria, p. 62–67. In Physical Murphy, Jr., and C. K. Grieshaber. 1994. In vivo-in vitro correlation of biochemistry. Prentiss-Hall, Englewood Cliffs, N.J. myelotoxicity of 9-methoxypyrazoloacridine (NSC-366140, PD115934) to 62. Wingard, J. R., A. D. Hess, R. K. Stuart, R. Saral, and W. H. Burns. 1983. myeloid and erythroid hematopoietic progenitors from human, murine, and Effect of several antiviral agents on human lymphocyte functions and mar- canine marrow. J. Natl. Cancer Inst. 86:273–280. row progenitor cell proliferation. Antimicrob. Agents Chemother. 23:593– 45. Parker, L. M., J. M. Lipton, N. Binder, E. L. Crawford, M. Kudisch, and 597. M. J. Levin. 1982. Effect of acyclovir and on human hematopoietic 63. Yarchoan, R., C. F. Perno, R. V. Thomas, R. W. Klecker, J.-P. Allain, R. J. progenitor cells. Antimicrob. Agents Chemother. 21:146–150. Wills, N. McAtee, M. A. Fischl, R. Dubinsky, M. C. McNeely, H. Mitsuya, 46. Physician’s Desk Reference. 1994. Cytovene᭨ (ganciclovir), p. 2357–2361. J. M. Pluda, T. J. Lawley, M. Leuther, B. Safai, J. M. Collins, C. E. Myers, Physician’s desk reference, vol. 48. Medical Economics Co. Inc., Oradell, and S. Broder. 1988. Phase I studies of 2Ј,3Ј-dideoxycytidine in severe human N.J. immunodeficiency virus infection as a single agent and alternating with 47. Physician’s Desk Reference. 1994. Retrovir᭨ (zidovudine), p. 742–746. Phy- zidovudine (AZT). Lancet i:76–80. sician’s desk reference, vol. 48. Medical Economics Co. Inc., Oradell, N.J. 64. Youle, M. S., B. G. Gazzard, M. A. Johnson, D. A. Cooper, J. F. Hoy, H. 48. Pluda, J. M., T. P. Cooley, J. S. G. Montaner, L. E. Shay, N. E. Reinhalter, Busch, B. Ruf, P. D. Griffiths, S. L. Stephenson, M. Dancox, and A. R. Bell S. N. Warthan, J. Ruedy, H. M. Hirst, C. A. Vicary, J. B. Quinn, G. J. Yuen, on behalf of the European-Australian Acyclovir Study Group. 1994. Effects M. A. Wainberg, M. Rubin, and R. Yarchoan. 1995. A phase I/II study of of high-dose oral acyclovir on herpesvirus disease and survival in patients 2Јdeoxy-3Јthiacytidine (Lamivudine) in patients with advance human immu- with advanced HIV disease: a double-blind, placebo-controlled study. AIDS nodeficiency virus infection. J. Infect. Dis. 171:1438–1447. 8:641–649. 49. Richman, D. D., M. A. Fischl, M. H. Grieco, M. S. Gottlieb, P. A. Volberding, 65. Zhu, Z., M. J. M. Hitchcock, and J.-P. Sommadossi. 1991. and O. L. Laskin, J. M. Leedom, J. E. Groopman, D. Mildvan, M. S. Hirsch, DNA interaction of 2Ј,3Ј-didehydro-2Ј,3Ј-dideoxythymidine in human bone G. G. Jackson, D. T. Durack, S. Nusinoff-Lehrman, and the AZT Collabo- marrow cells. Mol. Pharmacol. 40:838–845.