Parallel genome analysis by one- and two-dimensional DNA fingerprinting in human gliomas

Changes in methylation patterns identified by two-dimensional DNA fingerprinting

Two-dimensional DNA fingerprinting (2-D fingerprinting) is a sensitive tool for genomic difference analysis between tumor DNA and constitutive DNA of glioma patients. Numerous differences were found even in low-grade gliomas. They can be interpreted as deletions, amplifications, rearrangements, HaeIII restriction site mutations, tandem repeat instabilities, or methylation differences. The influence of methyl groups on the melting behavior of double-stranded DNA fragments in a denaturing gradient gel was demonstrated by analyzing the migration of lambda-phage DNA fragments in 2-D fingerprint gels. A characteristic intensity shift between two neighboring spots in several glioma samples was identified and verified by rehybridization of 2-D filters with a cloned DNA fragment corresponding to the lower spot in 10 out of 11 pilocytic astrocytomas. We hypothesized that this shift may be related to an alteration in the methylation pattern of the tumor DNA. This was specifically tested by analyzing the underlying 750 bp genomic fragment (including 21 CpG dinucleotides) with bisulfite treatment of agarose-embedded DNA. A methylation grade of 88% in tumor DNA as compared to 96% in blood DNA was found. Although only one CpG is located in the melting domain of the cloned fragment, this particular CpG is methylated in all blood samples, but mostly demethylated in the tumor samples. In conclusion, we demonstrate that 2-D fingerprinting may be a powerful tool for the detection of DNA methylation changes in genomic difference analysis.

Genomic difference analysis by two-dimensional DNA fingerprinting reveals typical changes in human low-grade gliomas

Cytogenetic and molecular analyses such as allelotyping studies have revealed several genetic changes typical for human glial neoplasms. However, most studies to date have involved malignant gliomas and thus are likely to reflect late events of tumor progression. To elucidate the initial events of glial tumor growth, we performed a genome-wide search for genetic alterations in the DNA of 43 low-grade gliomas as compared to the constitutional DNA of the patients' peripheral blood leucocytes using the two-dimensional (2D) DNA fingerprint approach. Reliable results were obtained for 28 blood/tumor sample pairs (13 astrocytomas, 9 pilocytic astrocytomas, 1 oligodendroglioma, 3 oligoastrocytomas, and 2 ependymomas). DNA was digested with the restriction enzyme HaeIII and the resulting fragments were separated on 2D gels according to size and sequence in the first and second dimensions, respectively. Patterns of hundreds of spots were generated by hybridization with four different mini- and microsatellite core probes. A total of 655 to 1,122 spots could be visualized per sample. Comparison of blood and tumor spot patterns revealed two to 11 reproducible changes per patient. Most of the differences were spot losses (77.1%), while the others appeared to be gains or amplifications. Exactly the same changes were found in tumor recurrences which lacked histological signs of progression. When comparing different patients, many of the affected spots tended to cluster in particular areas of the gel as revealed by computer-aided comparison of all spot patterns. Eleven different spot clusters were identified which may correspond to several major deletion targets. This study provides the basis for the future molecular cloning of the candidate tumor suppressor genes affected by the common spot losses and will allow new insights into the genetic mechanisms of glial tumorigenesis.

A novel standardization method for two-dimensional DNA fingerprints

Two-dimensional (2-D) DNA fingerprinting is a technique that allows for parallel genome analysis through the simultaneous detection of up to 500 mini- or microsatellite loci on a 2-D gel. Separation is performed according to size and melting temperature in the gel. In the application of this technique in genome analysis, a standardized method for the identification of individual spots is required. However, due to the polymorphic nature of up to 80% of the spots, existing standardization methods that have been primarily developed for 2-D protein patterns are not suitable for this task. We developed a robust method that standardizes 2-D DNA fingerprint spots on the basis of melting temperature - or denaturing gradient position - and fragment size. An external marker was used as a basis for standardization. A normalization surface was calculated over the gel dimensions by adapting an established numerical iteration technique previously used in physics termed "relaxation method". The relaxation method works robustly with the irregularly spaced marker spots. The evaluation of the method for a spot of preknown position derived from the TP53 gene revealed a median observed error below 1% for fragment length and denaturing gradient position. The search for candidate minisatellite loci in genomic difference analysis depends on the reliable identification of alleles of this locus in different individuals. We proved experimentally that alleles of a single minisatellite locus cloned from a 2-D gel cluster on an isothermal line can be reliably identified using the presented standardization method. In conclusion, a standardization tool for a broader application of 2-D DNA fingerprinting in both tumor analysis and possibly parallel mutation screening is now available.

Cloning of minisatellite-containing sequences from two-dimensional DNA fingerprinting gels reveals the identity of genomic alterations in low-grade gliomas of different patients

Two-dimensional (2-D) DNA fingerprinting was used to investigate genomic changes in human low-grade gliomas of different subtypes. DNA variations were identified in the 2-D hybridization patterns as spot losses or gains. Computer-aided matching of spot patterns from different patients revealed a clustering of spot changes at particular areas in the gel. Representative spots of each cluster were cloned using a spot cloning protocol which includes the preparation of a duplicate and a master gel. The DNA fragments of the 2-D gels were transferred to DEAE and nylon membrane, respectively. After hybridization of the master blot with a minisatellite core probe, the position of a particular spot was determined with reference to the lambda DNA fragments used as external markers in both gels. The gel spot DNA was recovered from the DEAE membrane by high salt elution and was polymerase chain reaction (PCR)-amplified after ligation of adaptor oligo cassettes. The PCR products were cloned and used as locus-specific probe for the rehybridization of the 2-D blots. One of these probes detected a spot loss in 7 of 28 low-grade gliomas of different subtypes analyzed. Another probe revealed a characteristic intensity shift in 8 of 9 pilocytic astrocytomas between two neighboring spots. The target sequence of this highly specific effect was assigned to chromosome 11q14 by in situ hybridization of a P1 clone harboring the affected genomic region. Thus, we successfully established a spot cloning procedure for the generation of locus-specific probes that may be instrumental in the discovery of the critical early events of glioma pathogenesis.

Marker pattern instabilities as a major cause of reproducibility problems in two-dimensional DNA fingerprinting

Two-dimensional (2-D) DNA fingerprinting is a promising technique for multilocus analysis of eukaryotic genomes. It has been successfully applied to the detection of DNA variation in tumors, to linkage analyses and to genomic comparisons of inbred mouse strains. However, there are still problems with inter-gel comparisons of 2-D DNA typing patterns as documented by the inter-gel reproducibility rates reported in the literature, which range from 84 to 98%. The basis for standardization in almost all of these studies has been a set of lambda fragments (digested separately with the restriction enzymes HaeIII, RsaI, Bg/I) that produces a spot pattern scattered across the gel. These spots are used as markers for gel comparisons. Since we noticed considerable variations in the marker spot patterns, we evaluated the properties of the lambda marker using both computer simulation and an empirical analysis of forty independent consecutive gels from our laboratory. We explain the instabilities of the spot pattern on the basis of the melting properties of the individual lambda fragments. A subset of spots is presented that has been stable in all our experiments. Only this set of spots should be used for gel standardization purposes until a new, completely reproducible marker becomes available. Finally, suggestions for an improved marker system are made.