Pyrrolo[2,3-d]pyrimidine nucleosides as inhibitors of human cytomegalovirus

Inhibition of human cytomegalovirus in culture by alkenyl guanine analogs of the thiazolo[4,5-d]pyrimidine ring system

A series of alkyl and alkenyl guanine analogs containing a thiazolo[4,5-d]pyrimidine ring system were prepared by reaction of the appropriate alkyl halide with the sodium salt of the heterocycle. In preliminary antiviral efficacy evaluations against laboratory strains of both human cytomegalovirus (HCMV) and herpes simplex virus types 1 and 2, it was determined that two of the compounds (T70072 and T01132) were more active and less toxic in stationary-phase cell monolayers than were the other derivatives tested. T01132 and T70072, which have 2-pentenyl and 3-methyl-2-butenyl moieties attached to position 3 of the 5-aminothiazolo[4,5-d]pyrimidine-2,7-dione, respectively, were then more extensively evaluated for anti-HCMV activity. The concentrations of T01132 and T70072 required to inhibit HCMV by 50% in plaque reduction assays were approximately 0.5 and 6.8 microM, respectively. These two compounds inhibited the growth of KB, MRC-5, or Vero cells at concentrations of 75 to 150 microM, depending upon the cell line. In bone marrow progenitor cells T01132 was slightly less toxic than ganciclovir (DHPG). The 50% inhibitory concentrations of T01132 against clinical isolates and DHPG-resistant strains of HCMV were approximately the same as those obtained for laboratory strains of HCMV (approximately 0.5 microM). When tested in combination with DHPG, the resultant antiviral activity was determined to be additive but not synergistic. Experiments performed using variations of the viral multiplicity of infection (MOI) demonstrated that T01132 was more active than DHPG at a low MOI (0.002 or 0.02). However, when a higher MOI (0.2 or 2.0) was used, DHPG was more efficacious than T01132. In experiments in which drug was added at various times post-viral infection, T01132 was most effective when added within the first 24 h post-HCMV infection while DHPG was able to protect cells in this assay system when added up to 48 h postinfection, indicating that T01132 is exerting its antiviral effect on events leading up to and possibly including viral DNA synthesis. The data presented in this report suggest that the antiviral activity of alkenyl-substituted thiazolopyrimidine derivatives may represent a mechanism of action against herpesviruses alternative to that of classical nucleoside analogs such as acyclovir or DHPG.

Relationship between cytotoxicity and conversion of thiosangivamycin analogs to toyocamycin analogs in cell culture medium

Non-nucleoside analogs of the pyrrolopyrimidine nucleosides toyocamycin, sangivamycin and thiosangivamycin have been synthesized and their cytotoxicity in mammalian cells determined. While studying the effects of 5-thioamide-substituted analogs on cell growth, we observed an interesting phenomenon in which cells recovered spontaneously from growth inhibition during extended incubations. HPLC studies demonstrated that the 5-thioamide moiety of several structurally dissimilar 7-substituted 4-aminopyrrolo[2,3-d]pyrimidines, including thiosangivamycin, is unstable in cell culture medium and is converted to the corresponding 5-nitrile with a half-life of approximately 48 h. In contrast, different substituents at the 4-position of the heterocycle significantly affected the stability of the 5-thioamide moiety. Conversion of the thioamide to the nitrile was caused by components in the cell culture medium, not components of serum. The above observations demonstrate that caution should be exercised in interpreting biological data obtained in vitro for 5-thioamide pyrrolo[2,3-d]pyrimidines.

Activity of triciribine and triciribine-5'-monophosphate against human immunodeficiency virus types 1 and 2

Triciribine (TCN) and its 5'-monophosphate (TCN-P) are novel tricyclic compounds with known antitumor activity; TCN-P is currently in phase II human clinical trials. We now report that these compounds have potent and selective activity against HIV-1 and HIV-2. Using a syncytial plaque assay, TCN and TCN-P were active against HIV-1 at 0.01-0.02 microM and had differential selectivities of 2250 and 1900, respectively, compared to 1850 for AZT. In contrast, TCN and TCN-P had minimal selectivity against human cytomegalovirus (50 and 27, respectively). TCN and TCN-P markedly inhibited HIV-1-induced p24 core antigen production, reverse transcriptase, and infectious virus production in a dose-dependent manner using HIV-1 acutely infected CEM-SS, H9, and persistently infected H9IIIB and U1 cells. In acutely infected PBL cells, TCN and TCN-P inhibited reverse transcriptase and infectious virus production but not p24 core antigen production. Using a microtiter XTT assay, TCN and TCN-P were active against a panel of HIV-1 and HIV-2 strains at IC50 values ranging from 0.02 to 0.46 microM. Evaluation of matched pairs of predrug and postdrug therapy HIV-1 isolates established that AZT-resistant and TIBO-resistant variants of HIV-1 were sensitive to TCN or TCN-P. Furthermore, unlike AZT and other fraudulent nucleosides, neither TCN, TCN-P, nor TCN-TP inhibited the viral reverse transcriptase. Thus, even though triciribine is a nucleoside chemically, it does not act biologically by classic nucleoside modalities but rather by a unique mechanism yet to be elucidated.

Improved synthesis and biological evaluation of an acyclic thiosangivamycin active against human cytomegalovirus

We previously described the synthesis and in vitro antiviral activity of an acyclic thiosangivamycin analog (Gupta et al., 1989a). In order to extend these initial studies, a new, multi-gram synthesis of 4-amino-7-[(2-hydroxy- ethoxy)methyl]pyrrolo]2,3-d]pyrimidine-5-thiocarboxamide (compound 229) was achieved in 5 steps from the known 5-amino-2-bromo-3,4-dicyanopyrrole in good overall yield. In plaque reduction assays with HCMV, compound 229 had an IC50 of 7 microM; in yield reduction assays the IC90 was 25 microM. The compound was less active against MCMV, HSV-1, HSV-2, and least active against VZV. Concentrations of compound 229 up to 32 microM did not affect the growth of KB cells for incubation periods up to 72 h. At 100 microM, a prolongation in population doubling time from 21 h (untreated) to 35 h was noted. This inhibition, however, was reversible upon removal of the compound suggesting the inhibition was cytostatic rather than cytotoxic. Flow cytometric studies with compound 229 in HFF cells revealed an accumulation of cells in S phase and a concurrent loss of cells in G2/M phase, suggesting an early S phase blockage. We conclude there is adequate separation between antiviral activity and cytotoxicity to merit further work with this class of pyrrolopyrimidines.

Activity of acyclic halogenated tubercidin analogs against human cytomegalovirus and in uninfected cells

Novel acyclic halogenated tubercidins (4-amino-5-halo-7-[(2-hydroxyethoxy)-methyl]pyrrolo[2,3-d]pyrimidines) were examined for their ability to inhibit human cytomegalovirus (HCMV) in yield reduction assays. 5-Bromo acyclic tubercidin (compound 102) was a more potent inhibitor of virus replication than the chloro- and iodo-substituted analogs (compounds 100 and 104). At a 100 microM concentration, the bromo and chloro compounds were more potent than acyclovir but not ganciclovir. Virus titers were reduced more than 99% by compounds 102 and 104 whereas compound 100 and the equally potent acyclovir reduced titers by only 90%. Quantitation of viral DNA by DNA hybridization demonstrated strong inhibition of HCMV DNA synthesis by these compounds. The most potent inhibitor, compound 102, had a 50% inhibitory (I50) concentration (1.6 microM) comparable to that of ganciclovir (1.8 microM). Cytotoxicity in uninfected human cells was evaluated and revealed the following: cell growth rates slowed markedly in the presence of 10 microM compound 102 whereas the same concentration of compounds 100 and 104 led to only a slight prolongation of population doubling time; these compounds inhibited cellular DNA synthesis but not RNA or protein synthesis, as measured by incorporation of radiolabeled precursors into acid-precipitable macromolecules; flow cytometry indicated that compound 102 was a mid-S phase blocker, and adenosine antagonized the inhibition of [3H]dThd incorporation by compound 102. Together, these results demonstrate that compound 102 is a potent and selective inhibitor of viral and cellular DNA synthesis and that acyclic halogenated pyrrolo-pyrimidine nucleosides may have therapeutic potential.

Inhibition of cell growth and DNA, RNA, and protein synthesis in vitro by fentanyl, sufentanil, and opiate analgesics

We have studied the cytotoxic nature of two groups of narcotic analgesics. Group 1 consists of the opioids, morphine, codeine, hydromorphone, thebaine, and etorphine. Group II contains but two phenylpiperidine-type narcotics, fentanyl and sufentanil. To measure cytotoxicity, three different bioassays were employed using an established line of human cells. Specifically, the effects of narcotic analgesics on DNA, RNA, and protein synthesis were measured by following the uptake and incorporation of radiolabeled thymidine, uridine, and amino acids, respectively. Inhibition of cell growth also was studied by measuring population doubling times of logarithmically growing cells in the presence (or absence) of the test compounds. Lastly, cloning efficiencies of cells were determined in the presence of both groups of compounds. Group I compounds were significantly less inhibitory than Group II compounds by all three bioassays. Moreover, flow cytometric DNA analysis of cells treated with 100 and 320 microM etorphine HCl showed essentially no effects on cell cycle distribution. These in vitro results thus suggest that (1) fentanyl and sufentanil are inherently more cytotoxic than the opioid narcotics in Group I, and (2) the highly potent morphinoid drug etorphine HC1 appears to have special promise as a transdermal narcotic to control pain.

A microtiter virus yield reduction assay for the evaluation of antiviral compounds against human cytomegalovirus and herpes simplex virus

Although the virus yield reduction assay is a powerful technique for evaluating the efficacy of antiviral compounds, it is not routinely utilized due to its labor-intensive nature. This procedure was modified, developed, thereby reducing greatly the time and effort required to perform yield reduction assays. Monolayer cultures of mammalian cells were grown in 96-well microtiter tissue culture plates and infected with virus. Test compounds were added and serially diluted directly with the plates. Following a cycle of virus replication, culture lysates were made and serially diluted in a separate set of uninfected cultures grown in microtiter plates. The cultures were incubated, plaques were enumerated in wells containing 5 to 20 plaques, and virus titers were calculated. To illustrate the use of the assay the known antiviral drugs acyclovir and ganciclovir were evaluated using this procedure. Ninety percent inhibitory concentrations for the respective drugs were 3 microM and 0.7 microM against herpes simplex virus type 1 and 60 microM and 1 microM against human cytomegalovirus.

Comparison of in vitro biological properties and mouse toxicities of three thymidine analogs active against human immunodeficiency virus

Three analogs of thymidine, D4T [2',3'-didehydro-2',3'-dideoxythymidine; 1-(2,3-dideoxy-beta-D-glyceropent-2-enofuranosyl)thymine], FddT (3'-fluoro-3'-deoxythymidine), and AZT (3'-azido-3'-deoxythymidine), were compared in biological tests designed to assess their potential utility as anti-human immunodeficiency virus (HIV) agents. The in vitro potencies of these compounds against HIV infection in CEM cells were measured, with FddT and AZT being more potent than D4T. The cytotoxicities of D4T, FddT, and AZT for CEM cells were comparable. The triphosphates of these three derivatives inhibited purified HIV reverse transcriptase, and their affinities for this polymerase were found to be 1 or 2 orders of magnitude greater than that for the normal substrate, dTTP. D4T was less toxic than FddT or AZT for cultured human and mouse bone marrow cells (granulocyte-macrophage CFU). The three compounds had similar toxicities for human progenitor erythrocyte burst-forming units. In a 30-day mouse toxicity study, AZT and FddT produced a similar spectrum of hematopoietic toxicities. These toxic effects occurred at much lower doses of FddT than of AZT. At the higher doses of FddT, a significant incidence of lethality occurred. By contrast, D4T was considerably less toxic than both AZT and FddT in this study. The dose-limiting toxicity of D4T in mice was hepatotoxicity. The very different phosphorylation patterns of D4T, its lower toxicity, and its comparable potency relative to FddT and AZT suggest that the potential of D4T as an anti-HIV agent should be further explored.

Flow cytometric evaluation of the cytotoxicity of novel antiviral compounds

Two acyclic analogs of bromotubercidin were tested for cytotoxic effects on uninfected cells by monitoring cell growth and measuring cell cycle perturbations using flow cytometry. As reported elsewhere, 5-bromotubercidin analogs in which ribose was replaced by 2-hydroxyethoxymethyl (compound 102) or by 1,3-dihydroxypropoxymethyl (compound 183) were potent inhibitors of human cytomegalovirus (HCMV) replication in vitro (Pudlo et al.: Journal of Medicinal Chemistry 31:2086-2092, 1988). Because these compounds also inhibited the growth of uninfected cells, we performed kinetic studies with an established neoplastic line of human cells (KB) using flow cytometry. Growth of KB cells treated with either compound 102 or 183 were inhibited in a dose-dependent manner. Growth inhibition by compound 183, however, was not fully expressed for at least 24 h. DNA analysis by flow cytometry showed that a 4-h incubation with 10 microM compound 102 caused a decrease of cells in G2/M phase. Cells began to accumulate in early S phase by 12 h of incubation, leading to mid S phase accumulation at 21 h. Compound 183 at 10 microM slightly decreased the number of cells in G2/M phase after a 4-h incubation, and led to accumulation of DNA in S phase after a 12-h incubation. By 24 and 30 h, DNA histograms appeared similar to those of control cells but with a slight accumulation of the population in early S phase. In separate experiments, drugs were removed following a 24-h incubation. After removal of compound 102, KB cell growth resumed with a normal population doubling time. In contrast, the effects of compound 183 were not reversible, suggesting the two compounds acted by different biochemical mechanisms.

Herpes simplex virus type 1 ribonucleotide reductase null mutants induce lesions in guinea pigs

Two herpes simplex virus type 1 ribonucleotide reductase null mutants, hrR3 and ICP6 delta, produced cutaneous lesions in guinea pigs as severe as those of wild-type strains. The lesions induced by hrR3 resulted from in vivo replication of the mutant virus, suggesting that this virus-encoded enzyme is nonessential for virus replication in guinea pigs.

Antagonism of the cytotoxic but not antiviral effects of ara-sangivamycin by adenosine

Inhibition of DNA synthesis by ara-sangivamycin was antagonized by adenosine. The 50% inhibitory concentrations increased 1.6- to 32-fold in the presence of 1.0 to 50 microM adenosine, respectively. In contrast, the inhibition of human cytomegalovirus replication by ara-sangivamycin was not antagonized by as much as 50 microM adenosine. This suggests that different enzymes were responsible for the phosphorylation of ara-sangivamycin in uninfected and infected cells.

Topical delivery of liposomally encapsulated interferon evaluated in a cutaneous herpes guinea pig model

The topical delivery of liposomally encapsulated interferon was evaluated in the cutaneous herpes simplex virus guinea pig model. Application of liposomally entrapped interferon caused a reduction of lesion scores, whereas application of interferon formulated as a solution or as an emulsion was ineffective. The method of liposomal preparation rather than the lipid composition of the bilayers appeared to be the most important factor for reducing lesion scores. Only liposomes prepared by the dehydration-rehydration method were effective. This finding implied that the dehydration and subsequent rehydration of the liposomes facilitate partitioning of the interferon into liposomal bilayers, where the drug is positioned for transfer into the lipid compartment of the stratum corneum. Liposomes do not appear to function as permeation enhancers but seem to provide the needed physicochemical environment for transfer of interferon into the skin.

Acquisition of cytomegalovirus infection: an update

Human cytomegalovirus (CMV) is a ubiquitous deoxyribonucleic acid virus that commonly infects a majority of individuals at some time during their life. Although most of these CMV infections are asymptomatic, certain patient groups are at risk to develop serious illness. Understanding the epidemiology of this virus is a key element in the development of strategies for preventing CMV disease. However, a number of features of this virus complicate such understanding. Following infection, CMV can remain latent, with subsequent reactivation; the factors controlling latency and reactivation and those factors which determine whether a CMV infection will be symptomatic are unknown. CMV disease can be acquired by natural routes, including horizontal and vertical transmission. Due to the ubiquity of CMV, the delineation of CMV transmission by these natural routes is complicated by the myriad of possible sources. Moreover, concerns over the risk of CMV transmission to the seronegative pregnant female have been raised in relation to preventing CMV transmission. By using molecular biologic techniques, much knowledge has been gained regarding the transmission of CMV disease by natural routes; however, a number of questions remain unanswered. The transmission of CMV infection by natural routes is therefore reviewed and the issues are highlighted. Primary infection, reactivation, and reinfection are the types of active CMV infections that can occur in an immunocompromised patient. In addition to natural routes of infection, introduction of presumably latently infected organs and requirements for multiple blood transfusions increase potential exposure to CMV in the immunocompromised patient. Understanding the epidemiology of CMV infections in the immunocompromised patient is difficult and in some instances controversial due to the complexity and interdependency of a number of factors which lead to CMV infection. In an immunocompromised individual, a major risk factor in developing overt CMV-related disease is associated with the serological status of an organ donor, the recipient, and the blood product given to these patients. In addition, a large body of inferential data supports the transmission of CMV by blood products or organs from seropositive donors; however, the mechanisms by which transmission occurs remain unclear. The possible sources and mechanisms of transmission of CMV infections in the immunocompromised host are reviewed. Lastly, strategies for the ultimate prevention of CMV disease are discussed in light of the epidemiology of CMV infections. To date, these strategies have included use of CMV-seronegative blood products or organs, antiviral agents, and vaccines.(ABSTRACT TRUNCATED AT 400 WORDS)

Antiviral activities of 2'-deoxyribofuranosyl and arabinofuranosyl analogs of sangivamycin against retro- and DNA viruses

Eight sugar-modified pyrrolopyrimidine nucleoside analogs related to the antibiotic sangivamycin were evaluated in cell culture against herpes simplex types 1 (HSV-1) and 2 (HSV-2), cytomegalovirus (CMV), adenovirus, and visna virus. Five of the compounds were highly active against most of the viruses with 50% inhibition (ED50) values of 1-10 microM. The selectivity of the agents was low, with inhibition of uninfected cell proliferation occurring within 5-fold that of the virus ED50 for most of the viruses. The compounds did not possess RNA virus-inhibitory activity when evaluated against certain myxo-, paramyxo-, picorna-, reo-, rhabdo-, and togaviruses. Two of the nucleosides were tested further in a cell line persistently infected with Friend leukemia virus where they were inhibitory to both virus yield and cell proliferation at 4-5 microM. Several of the sangivamycin analogs were tested in animal models using a twice-a-day treatment regimen. They proved to be inactive against HSV-1, murine CMV and/or Friend leukemia virus infections in mice.

(S)-1-(3-hydroxy-2-phosphonylmethoxypropyl)cytosine, a potent and selective inhibitor of human cytomegalovirus replication

From a series of phosphonylmethoxyalkylpurine and -pyrimidine derivatives, (S)-1-(3-hydroxy-2-phosphonylmethoxypropyl)cytosine [(S)-HPMPC] emerged as a particularly potent and selective inhibitor of the replication of human cytomegalovirus (CMV). Its potency against CMV was similar to that of the structurally related adenine derivative (S)-HPMPA but higher than that of the reference compounds phosphonoformate and 9-(1,3-dihydroxy-2-propoxymethyl)guanine (DHPG). The minimum concentrations of phosphonoformate, DHPG, (S)-HPMPA, and (S)-HPMPC required to inhibit CMV plaque formation by 50% were 15, 0.7, 0.1, and 0.07 microgram/ml, respectively. The selectivity indices of phosphonoformate, DHPG, (S)-HPMPA, and (S)-HPMPC, as determined by the ratio of the 50% inhibitory concentration for cell growth to the 50% inhibitory concentration for plaque formation for CMV (AD-169 strain), were 14, 150, 200 and 1,500, respectively. Corresponding values for the CMV Davis strain were 20, 200, 100, and 1,000, respectively. (S)-HPMPC was inhibitory to CMV plaque formation even when added to the cells at 24 or 48 h postinfection. When (S)-HPMPC was added immediately postinfection, a 24- or 48-h incubation time sufficed to obtain a marked inhibitory effect on CMV replication. Such limited incubation time was insufficient for DHPG to achieve any protection against CMV.

Inhibition of herpes simplex virus DNA replication by ara-tubercidin

Preliminary studies of the biochemical basis for the antiviral activity of the pyrrolo[2,3-d]pyrimidine nucleoside ara-tubercidin were conducted. Herpes simplex virus DNA synthesis was 3-fold more sensitive to inhibition by ara-tubercidin than was cellular DNA synthesis. Partially purified herpes DNA polymerases were more sensitive to inhibition by ara-tubercidin 5'-triphosphate than were cellular polymerases alpha and beta. Inhibition of viral DNA polymerase was competitive with dATP and noncompetitive with dTTP. The results suggest that the viral DNA polymerase plays a significant role in the antiviral activity of ara-tubercidin.