An in silico screen links gene expression signatures to drug response in glioblastoma stem cells.

Cancer stem cells (CSCs) are thought to promote resistance to chemotherapeutic drugs in glioblastoma, the most common and aggressive primary brain tumor. However, the use of high-throughput drug screens to discover novel small-molecule inhibitors for CSC has been hampered by their instability in long-term cell culture. We asked whether predictive models of drug response could be developed from gene expression signatures of established cell lines and applied to predict drug response in glioblastoma stem cells. Predictions for active compounds were confirmed both for 185 compounds in seven established glioma cell lines and 21 compounds in three glioblastoma stem cells. The use of established cell lines as a surrogate for drug response in CSC lines may enable the large-scale virtual screening of drug candidates that would otherwise be difficult or impossible to test directly in CSCs.

Update on NCI in vitro drug screen utilities.

Development of new anti-cancer drugs is a costly and risky proposition. The Developmental Therapeutics Program (DTP) of the National Cancer Institutes of the United States (U.S.) facilitates the drug development process by providing access to preclinical screening services. Since the early 1990's, DTP has screened tens of thousands of compounds against a panel of 60 human tumour cell lines representing nine tissue sites. At the same time, DTP began to accumulate information on the expression of molecular entities in the same 60 cell line panel. Many of these data are freely available to the public at. More recently, additional, more focused screens have entered the picture, with data also available through the web site. These include screening of roughly 100000 compounds against a panel of yeast mutants, and screening of the NCI Diversity Set in assays designed to detect effects on Molecular Targets of interest.

Thymidylate synthase as a molecular target for drug discovery using the National Cancer Institute's Anticancer Drug Screen.

Thymidylate synthase (TS) is a critical cellular target for cancer chemotherapeutics, particularly the fluoropyrimidine and antifolate classes of antineoplastic agents. One of the primary mechanisms of clinical insensitivity to these agents is through the overexpression of the target enzyme, TS. Thus, there is a need for the development of agents which selectively target TS-overexpressing malignant cells. To this end, we conducted a search for agents which potentially selectively target TS-overexpressing cells using two separate algorithms for identifying such compounds in the NCI Drug Repository by comparing cytotoxicity profiles of 30000 compounds with the TS expression levels measured by Western blot analysis in 53 cell lines. Using the traditional COMPARE analysis we were unable to identify compounds which maintain a selective ability to kill high TS-expressing cells in a subsequent four cell line validation assay. A new algorithm, termed COMPARE Effect Clusters analysis, enabled the identification of a particular drug cluster which contained compounds that maintained a selective ability to kill TS-overexpressing cell lines in the validation assay. While the identified compounds were selectively cytotoxic to TS-overexpressing cells, we found that they were not specifically targeting TS as a mechanism of action. Apparently, the overexpression of TS was providing a marker for sensitivity. This identified class of compounds which appears to be selectively cytotoxic against cells which overexpress TS may be useful for the development of therapeutics for those whose cancers overexpress TS de novo.

Dextran, gelatin, and hydroxyethyl starch do not affect permeability for albumin in cat skeletal muscle.

OBJECTIVE: To evaluate the effects of the three commercially available colloid solutions, 6% dextran 70, 6% hydroxyethyl starch (HES) 200/0.5, and 3.5% urea-linked gelatin on permeability for human albumin in a skeletal muscle in vivo model by evaluating their effects on the reflection coefficient for albumin. DESIGN: Controlled laboratory study. SETTING: University research laboratory. SUBJECTS: Eighteen adult cats. INTERVENTIONS: The autoperfused and denervated calf muscles of the cat hindlimb were placed in a plethysmograph. The transvascular fluid absorption induced by an increase in the colloid osmotic pressure following a fixed intravenous bolus of human albumin was analyzed, first before start of, and then during an intra-arterial infusion to, the muscle preparation of the synthetic colloid to be analyzed. Capillary filtration coefficient as a measure of microvascular fluid permeability (conductance) was analyzed before and after start of the synthetic colloid. MEASUREMENTS AND MAIN RESULTS: Arterial blood flow, arterial and venous blood pressures, total vascular resistance, tissue volume changes, capillary filtration coefficient, and plasma volume were measured before and during the colloid infusion. According to the Starling fluid equilibrium, the ratio between the reflection coefficients for albumin on two occasions (before and after infusion of the synthetic colloid) can be calculated from the maximum osmotic absorption rates induced by a fixed intravenous bolus infusion of albumin and from the capillary filtration coefficients. Obtained data were adjusted for different plasma volume at the two occasions. We found that none of the three synthetic colloids analyzed had any significant effect on the reflection coefficient for albumin. CONCLUSION: An effect on albumin microvascular permeability of the synthetic colloids dextran 70, HES 200/0.5, and urea-linked gelatin could not be shown by a method analyzing their effect on the reflection coefficient for albumin.

Effects on capillary fluid permeability and fluid exchange of albumin, dextran, gelatin, and hydroxyethyl starch in cat skeletal muscle.

OBJECTIVE: To evaluate the peripheral hemodynamic effects in a skeletal muscle in vivo model of the four commercially available colloid solutions, 20% human albumin, 6% dextran-70, 6% hydroxyethyl starch 200/0.5 (HES), and 3.5% urea-linked gelatin. DESIGN: Controlled laboratory study. SETTING: University research laboratory. SUBJECTS: Fifteen adult cats. INTERVENTIONS: The isolated, autoperfused, and denervated calf muscles of the cat hindlimb were placed in a plethysmograph. The colloids tested were given intra-arterially to the muscles in increasing rates of clinically relevant doses. MEASUREMENTS AND MAIN RESULTS: Arterial blood flow, arterial and venous blood pressures, total vascular resistance, tissue volume changes, and capillary filtration coefficient were measured before, during, and after the colloid infusions. The altered capillary filtration coefficient reflects a change in capillary fluid permeability. The capillary filtration coefficient was decreased by albumin and dextran, not affected by HES, and increased by urea-linked gelatin. Albumin induced transcapillary fluid absorption, gelatin induced transcapillary filtration, and no transcapillary fluid exchange was observed with dextran and HES. After discontinuation of the infusions, HES and gelatin induced a rebound transcapillary filtration. No such effect was seen after dextran and albumin. All colloids increased muscle blood flow. CONCLUSION: We conclude that capillary fluid permeability is decreased by albumin and dextran, unchanged by HES, and increased by gelatin. This and the differences in the rebound effect may contribute to the differences in the plasma volume expanding properties of the respective colloid. The increased blood flow induced by the colloids was more an effect of reduced vascular tone than of lowered blood viscosity.

Prostacyclin reduces microvascular fluid conductivity in cat skeletal muscle through opening of ATP-dependent potassium channels.

Prostacyclin is suggested to reduce microvascular permeability, but the cellular mechanisms mediating this response in the microvascular endothelial cells are still unknown. Considering that prostacyclin relaxes vascular smooth muscle cells via opening of ATP-dependent potassium channels, and opening of ATP-dependent potassium channels in the endothelial cells is suggested to influence microvascular permeability, this study was designed to test (1) if ATP-dependent potassium channels are involved in the regulation of microvascular hydraulic permeability, (2) if the permeability-reducing effect of prostacyclin is mediated through opening of ATP-dependent potassium channels, and (3) if cAMP is involved in this process. An autoperfused cat calf hindlimb was used as experimental model, and microvascular hydraulic permeability (conductivity) was estimated by a capillary filtration coefficient (CFC) technique. The potassium channel opener PCO-400 (0.5 microg x min(-1) per 100 g muscle, intra-arterially), prostacyclin (1 ng x min(-1) per kg body weight, intravenously) and the cAMP analogue dibutyryl-cAMP (24 microg x min(-1) per 100 g muscle, intra-arterially), decreased CFC to 77, 72 and 69% compared to control, respectively (p < 0.01). The decrease in CFC obtained by these substances was completely restituted after the start of a simultaneous infusion of the ATP-dependent potassium channel blocker glibenclamide (6 microg x min(-1) per 100 g muscle, intra-arterially; p < 0.01). Infusion of glibenclamide alone increased CFC to 107% of control (p < 0.05). In conclusion, the ATP-dependent potassium channels contribute to the regulation of microvascular hydraulic conductivity, and the prostacyclin permeability-reducing effect may act through this mechanism via increase in intracellular cAMP.

The chromosome bias of misincorporations during double-strand break repair is not altered in mismatch repair-defective strains of Saccharomyces cerevisiae.

Recombinational repair of a site-specific, double-strand DNA break (DSB) results in increased reversion frequency for nearby mutations. Although some models for DSB repair predict that newly synthesized DNA will be inherited equally by both the originally broken chromosome and the chromosome that served as a template, the DNA synthesis errors are almost exclusively found on the chromosome that had the original DSB (introduced by the HO endonuclease). To determine whether mismatch repair acts on the template chromosome in a directed fashion to restore mismatches to the initial sequence, these experiments were repeated in mismatch repair-defective (pms1, mlh1, and msh2) backgrounds. The results suggest that mismatch repair is not responsible for the observed bias.

A role for REV3 in mutagenesis during double-strand break repair in Saccharomyces cerevisiae.

Recombinational repair of double-strand breaks (DSBs), traditionally believed to be an error-free DNA repair pathway, was recently shown to increase the frequency of mutations in a nearby interval. The reversion rate of trp1 alleles (either nonsense or frameshift mutations) near an HO-endonuclease cleavage site is increased at least 100-fold among cells that have experienced an HO-mediated DSB. We report here that in strains deleted for rev3 this DSB-associated reversion of a nonsense mutation was greatly decreased. Thus REV3, which encodes a subunit of the translesion DNA polymerase zeta, was responsible for the majority of these base substitution errors near a DSB. However, rev3 strains showed no decrease in HO-stimulated recombination, implying that another DNA polymerase also functioned in recombinational repair of a DSB. Reversion of trp1 frameshift alleles near a DSB was not reduced in rev3 strains, indicating that another polymerase could act during DSB repair to make these frameshift errors. Analysis of spontaneous reversion in haploid strains suggested that Rev3p had a greater role in making point mutations than in frameshift mutations.

Chi enhances heteroduplex DNA levels during recombination.

The major pathway of homologous recombination in Escherichia coli, the RecBCD pathway, is stimulated by Chi sites. To determine whether Chi enhances an early or late step in recombination, we measured formation of heteroduplex DNA (hDNA) in extracts of lambda-infected E. coli. Chi elevated hDNA levels in these extracts, supporting a role for Chi early (before hDNA formation) in recombination. RecA protein and RecBCD enzyme were both necessary for detection of hDNA, indicating that they, too, act early. Analysis of a panel of recBCD mutants indicated that Chi-nicking activity was needed for Chi's stimulation of hDNA formation. These results support a previously proposed model of recombination. Further results suggested that RecBCD enzyme has an additional role late in recombination.

HLA DQ beta 3.2 identifies subtypes of DR4+ haplotypes permissive for IDDM.

The HLA class II-related susceptibility to type I insulin-dependent diabetes mellitus (IDDM) is examined in 94 multiplex families sorted by the presence or absence of a DR4+ haplotype in at least one diabetic family member. The families with DR4+ haplotypes are then sorted by the presence or absence of a DR4-linked DQ beta 3.2 allele. Further analysis assumes each multiplex family to represent a single diabetic genetic event and identifies the HLA class II haplotype(s) present in all affected members. The DQ beta 3.2 allele is present in over 95% of the multiplex families where DR4+ haplotypes segregate with IDDM, implying a major permissive role in determining susceptibility to IDDM.

HLA class II typing using oligonucleotide probes.

A DQ beta locus-specific oligonucleotide probe is used to identify DQ beta alleles present in six insulin-dependent diabetes mellitus (IDDM) multiplex families. Results of genotyping using this technique agree with those from cDNA probes and are often clearer. Allele-specific oligonucleotide probes for the DR beta locus are used to identify subtypes (Dw4 and Dw14) of the DR4 specificity in five IDDM multiplex families. This distinction cannot be made with available cDNA probes.

Molecular analysis of DQ beta 3.1 genes.

HLA class II genes have been implicated in susceptibility to a number of diseases. We have previously identified two allelic variants of DQw3 and have shown that DR4-DQ beta 3.2 haplotypes are associated with increased risk of IDDM whereas DR4-DQ beta 3.1 haplotypes are not. DR5 and DR8 DQw3+ individuals are exclusively DQ beta 3.1 and share numerous restriction sites within the DQ beta genes with DR4-DQ beta 3.1 individuals. In order to compare the DQ beta 3.1 genes associated with different haplotypes, we have sequenced coding and noncoding regions of the DQ beta genes from a DR4-DQ beta 3.1 HTC (ER) and a DR8-DQ beta 3.1 HTC (LUY). LUY and ER DQ beta genes share nucleotide substitutions in both the beta 1 and beta 2 exons, yielding six amino acid replacements distinguishing them from DQ beta 3.2. In the noncoding regions as well, LUY and ER share nucleotide substitutions distinguishing their DQ beta 3.1 genes from DQ beta 3.2. These data support the concept that the DQ beta 3.1 allele was introduced onto different backgrounds via homologous recombination.

Specific allelic variation among linked HLA class II genes.

Locus-specific oligonucleotides have been used as probes to detect polymorphic alleles for the HLA genes DQ beta, DQ alpha, DX alpha, and DO beta. These genes lie between the HLA DR and DP genes on chromosome 6, a region frequently implicated in intra-HLA recombination. In order to distinguish among these highly homologous HLA class II genes, we have identified sequences in the coding regions that are locus specific and have synthesized short oligonucleotide probes corresponding to these regions. We have used these probes in a gel hybridization procedure to analyze restriction-enzyme-digested genomic DNA and to assign polymorphic bands to a particular locus. Taq I-digested DNA hybridized with a DQ-beta-specific probe detects a single band per haplotype. The size of this band corresponds precisely to the expressed DQ beta gene and distinguishes among the serologically defined DQ alleles, providing a rapid method for genomic DQ typing. Similar analysis with an oligonucleotide probe specific for DX alpha, the nonexpressed alpha gene in the DQ subregion, detects 2 DX alleles, and hybridization with an oligonucleotide specific for the newly described DO beta gene detects 2 alleles at DO beta. Heterogeneity in linkage patterns among DQ, DX, and DO genes suggests that frequent recombinational events at multiple intergenic sites contributed to the generation of present-day haplotypes. One such recombinational event was identified directly in a family with serologically HLA-identical siblings, in which genomic analysis indicated a parental recombination event mapping between DR and DX, which correlated with unexpected alloreactivity.

Characterization of specific HLA-DQ alpha allospecificities by genomic, biochemical, and serologic analysis.

Bgl II restriction endonuclease digestion of genomic DNA from lymphoblastoid cell lines homozygous for HLA DR and DQ serological specificities, followed by hybridization with a DQ alpha cDNA probe, identified a genomic polymorphism characterized by two reciprocal patterns, one associated with DR 3, 5 and 8 and the other with DR 1, 2, 4, 7, and 9. The former pattern corresponded precisely to the reactivity of monoclonal antibody SFR20-DQ alpha 5, shown by Western blotting to react with isolated alpha-chains, but not with beta-chains. Additional variants of the DQ alpha genes were identified by using a locus-specific oligonucleotide probe for the DQ alpha gene, indicating differences among the DQ alpha 5-negative set of alleles. This analysis defines a set of DQ alpha allelic markers that are distinct from the well-established DQ serologic specificities DQw1, 2, 3 or "blank." Although most DQ alpha 5+ cells carry the DRw52 specificity associated with the DR beta 2 gene, analysis of DQ alpha polymorphisms on DR5, DQw1; DR8, DQw1; and DRw13, DQw1 cells verified that this DQ alpha family of alleles was not invariably linked to the DR beta 2 locus.

Identification of HLA-Dw14 genes in DR4+ rheumatoid arthritis.

Genes of the major histocompatibility complex, HLA, are associated with susceptibility to rheumatoid arthritis (RA), but the aetiology of this chronic inflammatory disease is not known. Synthetic oligonucleotide DNA probes were constructed to distinguish between two closely related but distinct alleles encoding the HLA-DR4 specificity in patients with RA. With these allele-specific oligonucleotide probes an uncommon DR4 genetic variant, Dw14, was identified in 6 of 7 RA patients homozygous for HLA-DR4. This allele may play an important part in susceptibility to RA.

Specific genomic markers for the HLA-DQ subregion discriminate between DR4+ insulin-dependent diabetes mellitus and DR4+ seropositive juvenile rheumatoid arthritis.

HLA-DR4, Dw4-associated haplotypes associated with IDDM and JRA were compared using genomic DNA restriction fragment analysis to distinguish among DQ beta and alpha alleles linked to DR4. DQ beta polymorphisms that subdivide the HLA-DQw3 specificity into DQ3.1 and 3.2 alleles were identified. More than 90% of DR4+ IDDM patients express one of these alleles, DQ3.2; restriction enzyme mapping indicates that the presence of this allele also accounts for the genomic fragment patterns previously reported in IDDM. Furthermore, haplo-identical siblings of DQ3.2 IDDM patients also carry the DQ3.2 allele, regardless of clinical presentation. In contrast, DR4+ JRA patients show no allelic preference at DQ beta, implicating different HLA genetic contributions in these two DR4-associated diseases.

Exon-specific oligonucleotide probes localize HLA-DQ beta allelic polymorphisms.

The HLA genetic region consists of a large multigene complex which includes a number of highly homologous alpha and beta genes encoding class II polypeptides, clustered in three major loci, DP, DQ, and DR. Analysis of genomic polymorphisms at each of these loci is of considerable interest due to the role of particular structural polymorphisms in immune function, but this analysis has been hampered by difficulty in distinguishing between such highly homologous loci. We have identified locus-specific and exon-specific class II gene sequences in order to produce synthetic oligonucleotide probes which hybridize specifically to DQ beta genes. Two such oligonucleotide probes are described which are specific for the beta 1 and beta 2 exons of DQ (DC) beta, which identify DQ beta genes in digests of cellular DNA and which can be used to characterize restriction sites flanking the two oligonucleotide-specific regions. By sequentially hybridizing these probes in modified Southern analyses, we have been able to generate a tentative "restriction map" of a newly identified DQ beta allele from digests of total genomic DNA. This oligonucleotide mapping technique discriminates between two HLA-DQw3+ alleles, DQ3.1 and DQ3.2, permitting the recognition of structural polymorphisms with DQ beta which are highly associated with type I diabetes mellitus.

Identification of a polymorphic variant associated with HLA-DQw3 and characterized by specific restriction sites within the DQ beta-chain gene.

Restriction endonuclease digestion of genomic DNA from 24 lymphoblastoid cell lines homozygous for the HLA class II specificity DQw3, followed by hybridization with a DQ beta-chain cDNA probe, identified a genomic polymorphism with variable BamHI and HindIII recognition sites. This restriction fragment pattern was found for several haplotypes associated with the DQw3 specificity, including some haplotypes positive for the HLA-DR specificities DR4, DR5, DRw8, and DRw12. The variant fragment pattern corresponds precisely with the reactivity of a monoclonal antibody, A-10-83, previously shown to define a serologic split of DQw3. Serologic detection of the specific DQw3.1 genomic polymorphism indicated that the corresponding DQ beta-chain variants are expressed. This polymorphic restriction fragment pattern, then, represents a selective marker for DQ beta-chain genes that appear to define a DQ beta-chain-associated specificity, here called DQw3.1.

HLA-DR4-associated haplotypes are genotypically diverse within HLA.

Biochemical diversity among products of class II HLA genes has been observed in individuals who appear to be HLA-D and DR-identical by cellular and serologic typing. We used techniques of restriction enzyme fragment analysis by Southern blotting to analyze this diversity at the level of cellular DNA. A panel of 17 HLA-DR4 homozygous cell lines (HCL) were investigated by using cDNA probes homologous to DQ beta, DQ alpha, and DR beta genes. Each probe was hybridized to cellular DNA digested with a series of different restriction endonucleases. Polymorphisms were observed with the use of the enzymes Pst I, Hind III, and Bam HI: Hybridization of cellular DNA digested with Hind III and Pst I with the DQ beta probe revealed specific polymorphisms, as did hybridization of the Pst I digest with the DQ alpha cDNA probe and the Bam HI digest with the DR beta probe. The observed differences fall into two categories: first, considerable diversity was seen between HLA-DR4 HCL that represent different HLA-D-defined haplotypes; second, diversity was also observed among HCL of the same DR4-associated HLA-D cluster. In contrast to the DQ cDNA probes, hybridization with the DR beta probe revealed relatively limited polymorphism by using a panel of different restriction endonucleases. Thus, although there is a general pattern of polymorphic restriction enzyme fragments homologous to DQ probes within an HLA-D cluster, the pattern seen for any particular cell line was not sufficiently distinct to assign an HLA-D or DR specificity.