Search

The precise classification of tumors into relevant molecular subtypes will facilitate both future research and optimal treatment. Here, the Lund Taxonomy system for molecular classification of urothelial carcinoma was applied to two large and independent cohorts of non-muscle-invasive tumors. Of 752 tumors classified, close to 100% were of the luminal subtypes, 95% urothelial-like (Uro; UroA, UroB, or UroC) and 5% genomically unstable. The obtained subtype structure organized the tumors into groups with specific and coherent gene mutation, genomic, and clinical profiles. The intrasubtype variability in the largest group of tumors, UroA, was caused by infiltration and proliferation, not considered as cancer cell type-defining properties. Within the UroA subtype, a HOXB/late cell-cycle gene expression polarity was found, strongly associated with FGFR3, STAG2, and TP53 mutations, as well as with chromosome 9 losses. Kaplan-Meier analyses identified the genomically unstable subtype as a progression high-risk group, also valid in the subgroup of T1 tumors. Almost all progression events occurred within 12 months in this subtype. Also, a general progression gene signature was derived that identifies high- and low-risk tumors. All findings were demonstrated in two independent cohorts. The Lund Taxonomy system is applicable to both non-muscle- and muscle-invasive tumors and may be a useful biological framework for translational studies.

Background For muscle-invasive bladder cancer (MIBC), no tissue biomarkers are available for clinical use to predict response to neoadjuvant chemotherapy. Objective To investigate how molecular subtypes impact pathological response and survival in patients receiving preoperative cisplatin-based chemotherapy. Design, setting, and participants Classification of a retrospective cohort of 149 patients was performed by tumor transcriptomic profiling and immunostaining. A cohort treated with radical cystectomy alone and public data sets were used for comparison and external validation. Outcome measurements and statistical analysis Complete pathological response in the cystectomy specimen (ypT0N0) and survival were compared in predefined molecular subtypes. Differential gene expression and chemotherapy response were explored beyond molecular subtypes. Results and limitations Patients with genomically unstable (GU) and urothelial-like (Uro) tumors had higher proportions of complete pathological response (16/31 [52%] and 17/54 [31%]), versus five out of 24 (21%) with the basal/squamous (Ba/Sq) subtype following neoadjuvant chemotherapy and radical cystectomy. Molecular subtype was independently associated with improved survival for patients with GU tumors (hazard ratio [HR] 0.29, 95% confidence interval [CI]: 0.11–0.79) and UroC tumors (HR 0.37, 95% CI: 0.14–0.94) compared with Ba/Sq tumors, adjusting for clinical stage. In addition, expression of the gene coding for osteopontin (SPP1) showed a subtype-dependent effect on chemotherapy response. Conclusions Urothelial cancer of the luminal-like (GU and Uro) subtypes is more responsive to cisplatin-based neoadjuvant chemotherapy. A second-generation of subtype-specific biomarkers, for example, SPP1, may be a way forward to develop a more precision-based treatment approach for neoadjuvant chemotherapy in MIBC. Patient summary This study shows that tumor classification by gene expression profiling and molecular subtyping can identify patients who are more likely to benefit from chemotherapy before radical cystectomy for muscle-invasive bladder cancer. Together with other markers for response, molecular subtypes could have a role in selective administration of such chemotherapy.

Several groups have during past years produced molecular classification schemes for bladder cancer. Even though no consensus on how to define a subtype exists, one approach has been to base definitions on how tumours cluster according to their mRNA expression profiles. In many cases, obtained profiles, and thus class defining features, are affected by signals from non-tumour cells within the biopsy. To overcome this issue, we combined gene expression analyses with analyses of the actual tumour cells by extensive immunohistochemistry (IHC). By this approach we were able to define tumour cell phenotypes i.e., subtypes defined by features of the tumour cells only, and adjust mRNA-based algorithms accordingly. In the present investigation we address the non-luminal Basal/Squamous-like (Ba/Sq) and Small cell/Neuroendocrine-like (Sc/NE) categories of tumours defined by mRNA-based classification. We make use of IHC data for 15 proteins, all known to be instrumental for defining molecular subtypes of urothelial carcinoma. We show that the UroB type of tumours, frequently grouped together with Ba/Sq, are different from the Ba/Sq entity at several essential features and is a derivative of Urothelial-like tumours (Uro). We show that the Sc/NE tumours are similar to but represents extreme versions of Genomically Unstable (GU) tumours. We apply clustering to 423 cases representing all subtypes using IHC data for 14 proteins and show that the obtained grouping conforms well with the mRNA-based classification. This work describes in detail the molecular pathology of non-luminal RNA-based bladder cancer subtypes and highlight similarities/dissimilarities suggestive of origin.

In our study the Bladder Cancer Molecular Taxonomy Group collaborated to extend a first consensus report, addressing the need for a consensus molecular classification for muscle-invasive bladder cancer (MIBC) that would support basic research and clinical trials. We provide such a consensus classification and offer a single-sample classifier (http://cit.ligue-cancer.net:3838/apps/consensusMIBC_web).

Introduction To describe the genetic relationship between the primary and recurrent bladder tumors and to identify their origins, different models such as the field cancerization, seeding, and intraepithelial migration have been proposed. The aim of this study was to investigate the genetic relationship in syn- and metachronous bladder tumors using different types of data, and to highlight supportive evidence for the proposed models. Methods The exophytic part of primary and recurrence urothelial tumors were collected by cold-cup biopsy from patients undergoing transurethral resection at the Lund University Hospital, Sweden, between 2001 and 2007. In total, 49 tumors from 18 patients and their 18 matched normal samples obtained from peripheral blood were collected and hybridized on HumanCNV370-Duo Genotyping BeadChips (Illumina Inc.). Gene expression data were generated using Illumina HumanHT-12 Expression BeadChip. Tumors were classified according the Lund classification system. For 27 tumor samples and matched normal blood samples from 10 patients, a custom target captured panel was used to sequence 1697 cancer genes. Results We used the fact that patients with non-muscle invasive bladder tumors show frequent local recurrences and often have multiple tumors to study re-initiation of tumor growth from the same urothelium. By extensive genomic analyses we show that tumors from the same patient are clonal. We show that gross genomic chromosomal aberrations may be present then absent in a recurrent tumor. By analyses of incompatible changes i.e., genomic alterations that cannot be reversed, we show that almost all tumors from a single patient may show such changes, thus the tumors cannot have originated from each other. Conclusions As recurring tumors share both genomic alterations and driver gene mutations these must have been present in the urothelium in periods with no overt tumor growth. The presence of similar mutational and CN profiles combined with incompatible events suggests that the primary and the recurrent tumors have a shared ancestry but are not originating from each other. This supports a model that includes a growing and evolving field of urothelial cells that occasionally, and locally, produce bursts of cellular growth leading to overt tumors.

Introduction Large efforts have been made to divide bladder cancer into prognostic or predictive subtypes by unsupervised or semi-supervised clustering of mRNA data. To validate and make use of subtypes, classifiers must be constructed that accurately predicts subtypes in new samples. Variations of nearest centroid classification is commonly used, where each class is represented as a vector of mean fold changes of selected genes. The similarity to these vectors is then used to predict the class of a new sample. Using relative fold change values means that classification becomes related to both the composition of the training cohort and the cohort the new sample originates from. Without thorough data preprocessing, this can result in major classification errors. Optimally, we want to classify a new sample in isolation in a manner robust to preprocessing, sample purity, and mRNA measurement technology. Methods We evaluated single sample classification methods based on using sets of gene-pair rules (e.g. Gene A>Gene B indicates Subtype X), using the R packages AIMS and SwitchBox, and a custom iterative RandomForest (RF) approach, as well as the single sample centroid approach used by Kamoun et. el. These were tested on Lund, TCGA, and IMvigor210 data. Results Rule-based and centroid based single sample classifiers were able to achieve similar accuracy in predicting LundTaxonomy subtypes in the datasets. The single sample centroid classifier based on raw expression values suffered from poor separation between predictions (delta between best and second-best prediction), which becomes worse as more genes were included. Rule-based methods, particularly RF, had much more distinct subtype calls. The custom RF approach outperformed the published AIMS and SwitchBox methods. Conclusions Each single sample classifier method had strengths and limitations. The RF based approach emerged as a promising new method, being robust even when using very few genes, while also allowing for evaluation of class stability and putative incorrect reference labels during training. The main challenge is how to select which gene-rules to use in the classifier. An R package (“multiclassPairs”) for building SwitchBox and RandomForest classifier models is currently being prepared.

Introduction Several groups, including ours, have identified STAG2 as a new bladder tumor suppressor gene. STAG2 is part of cohesin, a complex involved in chromosome segregation, DNA repair and chromatin organization. A role for STAG2 in hematopoietic stem cell self-renewal and differentiation has been shown. The mechanisms through which STAG2 contributes to UBC have not been elucidated, but we proposed that they involve processes different from altered chromosome segregation. STAG2 inactivation is associated with a class of differentiated, luminal tumours significantly enriched in the PPARg pathway, a driver of urothelial differentiation. We have set out to determine whether STAG2 inactivation alters stemness and differentiation balance in UBC. Methods We have established 2D cultures of primary murine urothelial cells (NU-1), applied integrated ChIPseq and RNAseq data, and generated organoids from conditional Stag2 knockout mice. Results Stag2 loss of expression is significantly associated with the luminal subtype, both in NMIBC and MIBC - Spontaneously immortalized NU-1 cells undergo urothelial differentiation upon PPARg activation and EGFR inhibition - In differentiated NU-1 cells, STAG2 localizes predominantly at promoters and enhancers and is important for the transcriptional regulation of the associated genes - ChIPseq and co-IP data indicate that PPARg and STAG2 - but not STAG1 - are present in the same complex and cooperate in the regulation of tissue-specific transcription - Stag2 knock out urothelial cells show higher clonogenic potential - Concomitant STAG2 loss and urothelial damage lead to hyperplasia, suggesting a role in homeostatic regeneration Conclusions STAG2 loss of expression is associated with luminal tumors. Our genomic and biochemical data suggest that STAG2 cooperates with PPARg in regulating urothelial differentiation and its loss alters gene expression, priming cells to proliferate. ATAC-seq experiments will shed light on the effects of Stag2 inactivation on genome accessibility and regulation of gene expression in proliferating and differentiated urothelial cells. Organoids and 2D cultures derived from conditional Stag2 knockout mice provide a novel platform for functional analyses and suggest a possible role of Stag2 in stem cell homeostasis.

Background One third of patients with stage T1 urothelial carcinoma (UC) progress to muscle-invasive disease requiring radical surgery. Thus, reliable tools are needed for risk stratification of stage T1 UC. Objective To investigate the extent to which stratification of stage T1 tumours into previously described molecular pathologic UC subtypes can provide improved information on tumour progression. Design, setting, and participants A population-based cohort of 167 primary stage T1 UCs was characterised by immunohistochemistry and classified into the molecular subtypes urobasal (Uro, 32%), genomically unstable (GU, 58%), and squamous-cell-carcinoma–like (SCCL, 10%). Outcome measurements and statistical analysis Progression-free survival using univariate and multivariate models. Results and limitations Subtype classification was validated using nine additional markers with known subtype-specific expression. Analysis of mRNA expression of progression biomarkers revealed a strong association with molecular subtype. Kaplan-Meier analyses showed that the risk of progression was low for Uro tumours and high for GU/SCCL tumours. High progression risk scores were found only for GU/SCCL tumours. Clinical risk factors such as multifocality, concomitant carcinoma in situ, invasion depth, lymphovascular invasion, and high CD3 + lymphocyte infiltration were observed almost exclusively in GU/SCCL cases. Conclusions Molecular subtypes Uro, GU, and SCCL were identified in an independent population-based cohort of stage T1 UCs. Biomarkers and clinical risk factors for progression were associated with molecular subtype. Rapidly progressing T1 tumours were of subtype GU or SCCL and had either a high progression risk score or an elevated CD3 + cell count. Patient summary We show that classification of stage T1 urothelial carcinoma into molecular subtypes can improve the identification of patients with progressing tumours.

We recently defined molecular subtypes of urothelial carcinomas according to whole genome gene expression. Herein we describe molecular pathologic characterization of the subtypes using 20 genes and IHC of 237 tumors. In addition to differences in expression levels, the subtypes show important differences in stratification of protein expression. The selected genes included biological features central to bladder cancer biology, eg, cell cycle activity, cellular architecture, cell-cell interactions, and key receptor tyrosine kinases. We show that the urobasal (Uro) A subtype shares features with normal urothelium such as keratin 5 (KRT5), P-cadherin (P-Cad), and epidermal growth factor receptor (EGFR) expression confined to basal cells, and cell cycle activity (CCNB1) restricted to the tumor-stroma interface. In contrast, the squamous cell cancer-like (SCCL) subtype uniformly expresses KRT5, P-Cad, EGFR, KRT14, and cell cycle genes throughout the tumor parenchyma. The genomically unstable subtype shows proliferation throughout the tumor parenchyma and high ERBB2 and E-Cad expression but absence of KRT5, P-Cad, and EGFR expression. UroB tumors demonstrate features shared by both UroA and SCCL subtypes. A major transition in tumor progression seems to be loss of dependency of stromal interaction for proliferation. We present a simple IHC/histology-based classifier that is easy to implement as a standard pathologic evaluation to differentiate the three major subtypes: urobasal, genomically unstable, and SCCL. These three major subtypes exhibit important prognostic differences.

The most important event to predict for patients with non– muscle-invasive (NMI) bladder cancer is progression to muscle-invasive growth. European Organization for Research and Treatment of Cancer (EORTC) risk tables have been used and validated as a prognostic tool for both cancer recurrence and progression. The tables include tumor grade according to the World Health Organization (WHO) 1973 system, tumor stage, and presence/absence of carcinoma in situ (CIS), complemented by three clinical parameters: prior recurrence rate, tumor diameter, and number of tumors. In the European Association of Urology (EAU) guidelines for NMI bladder cancer from2011, gradingwith both the threetiered WHO 1973 and the two-tiered 2004 systems are recommended [1]. The argument in favor of WHO 2004 has been less interobserver variability. However, the threetieredWHO 1999 has more precise criteria for grading than WHO 1973, reduces the interobserver variability, and is used in 30% of pathology laboratories in Europe, in addition to theWHO2004 system [2]. TheWHO1999 andWHO2004 systems are completely congruent and WHO 1999 is easily translated to WHO 2004; grade 2 and grade 3 tumors in WHO 1999 constitute high-grade carcinoma in the WHO 2004 system [3]. As opposed to the WHO 2004 system, WHO1999 can also be used for EORTC risk tables. Themajor difference between the WHO 1999 and WHO 2004 grading systems is that WHO 1999 defines a more distinct highgrade (G3) group of tumors. In this article, we give

During the past decades, the complexity of the bladder cancer genome has become evident. Early cytogenetic studies identified several patterns of chromosomal changes, particularly the frequent loss of chromosome 9. The cytogenetic approach was replaced by molecular methods, such as comparative genome hybridization (CGH) and loss of heterozygosity (LOH) analyses that describe genomic changes at a molecular and higher resolution. With these methods, the full complexity of the bladder cancer genome has been better appreciated. Using CGH and LOH analyses, it also became apparent that premalignant lesions of the bladder, such as hyperplasia and dysplasia, as well as carcinoma in situ (CIS), showed genomic changes. Whole genome analyses showed that low stage, low grade tumors generally show fewer changes than tumors of higher stage and grade. In addition, several genomic alterations were shown to be highly specific for more aggressive and invasive tumors. Based on the general association between complex genomic changes and tumor behavior, several investigations have been directed towards the identification of prognostic genomic markers for urothelial cancer. A complicating factor in the analysis and understanding of bladder cancer genomic progression is that recurring and, hence, chronologically later tumors may show genomes less rearranged than preceding tumors. Furthermore, morphologically normal urothelium in patients with bladder cancer frequently show the same type of genomic alterations as the tumor proper. This makes an issue of to what extent information on genomic changes will produce reliable prognostic information when limited to the tumor proper.

Introduction To describe the origin and the genetic relationship between primary bladder tumours and their recurrences, different models were proposed, such as, field cancerization, seeding, and intraepithelial migration. This study aims to investigate the genetic relationship in syn- and metachronous bladder tumours using different types of data, and to evaluate the proposed models. Material and methods The exophytic part of primary and recurrent urothelial tumours were collected by cold-cup biopsies from patients undergoing transurethral resection at the Lund University Hospital, Sweden, between 2001 and 2007. In total, 51 tumours from 19 patients and their 19 matched peripheral blood samples were collected and hybridised on Illumina HumanCNV370-Duo Genotyping array. Copy number data was extracted from GenomeStudio 2 software, and normalised using transformed quantile normalisation (tQN). Mirrored B allele frequency values were segmented using BAFsegmentation. Mutation calling was performed using targeted cancer gene sequencing panel. Gene expression data was generated using Illumina HumanHT-12 expression array. Results and discussions A manual inspection of all copy number alterations (CNAs) profiles was performed. Normal cells content, tumour ploidy, and sub clonality were calculated, and the relationship between the primary and recurrent tumours were estimated. Preliminary results show that CNAs are early and stable events during the tumour evolution. Part of the syn- and metachronous bladder tumours shared similar mutational and CNAs profiles. However, even with similar profiles, some events such as loss of heterozygosity in the primary tumours, were not observed in the recurrent tumours, constituting incompatible events. Samples with mutations in TP53 or amplification in MDM2 showed higher number of break points in the genome. A further analysis will be performed using the available CNAs, expression, and mutational data. Conclusion The genetic relationship in syn- and metachronous bladder tumours was investigated using different data types providing a better understanding of bladder cancer evolution. The presence of similar mutational and CNAs profiles combined with incompatible events suggests that the primary and the recurrent tumours were generated separately from a common ancestor. These findings highlight a tumour evolution scenario of a clonal cancerized field.

Gross mitotic disturbances are often found in malignant tumours, but not until recently have the molecular causes and the genomic consequences of these abnormalities started to become known. One potential source of mitotic instability is chromosomes with dysfunctional telomeres, giving rise to a high rate of chromatin bridges at anaphase. These bridges could lead either to structural chromosome rearrangements through chromatin fragmentation or to whole-chromosome losses through kinetochore-spindle detachment. Statistical meta-analyses have recently revealed that tumours with high rates of anaphase bridging, such as ovarian, head and neck, and pancreatic carcinomas, are characterised by multimodal distributions of genomic imbalances, consistent with a dramatically increased rate of chromosome rearrangements. In contrast, tumours without gross cell division disturbances are characterised by a monotonously decreasing distribution of genomic changes. This distribution follows a power-law, best described by a preferential attachment model in which the tolerance for chromosomal changes increases steadily with tumour growth. Even though many common cancers, such as breast, colorectal, and renal cell carcinomas adhere to this simple power-law dynamics, the underlying molecular mechanisms remain elusive.

Lung carcinomas are cytogenetically highly complex. In spite of this, patterns of recurrent chromosome aberrations have emerged. Apart from the frequent loss of 3p, losses of 4q, 5q, 8p, 9p, 10q, 13q, and 17p are common and gains often include 1q, 3q, 5p, and 8q. In the present study, we retrieved all aberrant lung carcinoma karyotypes, in total 432 cases, from the Mitelman Database of Chromosome Aberrations in Cancer and identified the most frequent imbalances. Each case was then classified with respect to the presence or absence of these imbalances and the data were statistically analyzed by means of principal component analysis, multidimensional scaling, and hierarchical cluster analysis. The analyses suggest that lung cancer develops through three pathways, initiated by +7, 3p-, and +12, respectively, and that the 3p- pathway is dominated by losses and the +12 pathway by gains. Gain of chromosome 7 was shown to be both important in the 3p- pathway and also forming a group of tumors containing +7 and +20 (with few additional changes). The distribution of the number of imbalances per tumor indicated that the karyotypic evolution might pass through three different phases. Phase I is characterized by tumors with few changes and by well-separated 3p- and +12 pathways. Phase II cases have an increased number of imbalances and exhibit less distinct 3p- and +12 pathways. Phase III tumors are polyploid and highly complex. No marked differences between the karyotypic profiles were found among morphologic subtypes, suggesting that lung cancer morphology is independent of the particular cytogenetic pathway operating in the tumor cells.

Although a great deal of information has accumulated regarding the mechanisms underlying constitutional DNA rearrangements associated with inherited disorders, virtually nothing is known about the molecular processes involved in acquired neoplasia-associated chromosomal rearrangements. Isochromosome 17q, or "i(17q)," is one of the most common structural abnormalities observed in human neoplasms. We previously identified a breakpoint cluster region for i(17q) formation in 17p11.2 and hypothesized that genome architectural features could be responsible for this clustering. To address this hypothesis, we precisely mapped the i(17q) breakpoints in 11 patients with different hematologic malignancies and determined the genomic structure of the involved region. Our results reveal a complex genomic architecture in the i(17q) breakpoint cluster region, characterized by large ( approximately 38-49-kb), palindromic, low-copy repeats, strongly suggesting that somatic rearrangements are not random events but rather reflect susceptibilities due to the genomic structure.

Oliver Patschan a,y,*, Gottfrid Sjodahl a,y, Gunilla Chebil , Kristina Lovgren , Martin Lauss , Sigurdur Gudjonsson , Petter Kollberg , Pontus Eriksson , Mattias Aine , Marten Ferno , Fredrik Liedberg , Mattias Hoglund b Department of Urology, Clinical Sciences, Skane University Hospital, Lund University, Malmo, Sweden; Department of Oncology, Clinical Sciences, Skane University Hospital, Lund University, Lund, Sweden; Department of Urology, Helsingborg Hospital, Helsingborg, Sweden

Human cancers are characterized by chromosomal aberrations, and an increasing number of specific balanced rearrangements have been found among malignant hematologic disorders. Most solid tumors, however, exhibit a much more complex cytogenetic pattern. Although these chromosome changes show a nonrandom distribution, tumor-specific aberrations are uncommon, and the solid tumors often contain a large number of abnormalities and also display extensive cytogenetic variability. The high level of karyotypic complexity has made a systematic characterization of the chromosomal patterns difficult. In order to better understand the biological relevance of highly abnormal karyotypes in tumor cell populations, novel statistical strategies are needed. We have developed and adapted several methods that may be useful for the evaluation of general patterns of karyotypic complexity, including distribution analysis of cytogenetic imbalances, temporal analysis for time of occurrence of aberrations, and principal component analysis for reconstructing karyotypic pathways. By applying these methods on the chromosomal changes presently known, distinct subgroups have been identified among breast, kidney, bladder, colon, and brain tumors.

Structural rearrangements of chromosome 8 are frequently encountered in squamous cell carcinomas of the head and neck (HNSCC). These aberrations often affect the centromeric region, resulting in the formation of isochromosome i(8q) and whole arm translocations. Some tumors may display structural rearrangements of 8p23. To characterize further the localization of the breakpoints in such rearrangements, 12 HNSCC known to carry pericentromeric rearrangements of chromosome 8 and 8p23 abnormalities were investigated with fluorescence in situ hybridization (FISH) by the use of 15 YAC clones spanning 8p23 and 8p11 to 8q11. FISH confirmed that all, except one, aberrations cytogenetically interpreted to be i(8q) were true, monocentric i(8q). Similarly, all whole-arm translocations appeared as centric fusions. It could thus be concluded that the essential outcome of these rearrangements is genomic imbalances and not rearrangement of genes in the pericentromeric region. By the use of five YAC clones mapping to 8p23, different breakpoints at the molecular level were disclosed in cases with cytogenetically identical 8p23 rearrangements. An evaluation of the genomic imbalances detected in the present series revealed that overrepresentation of 8q material was present in 11 of the 12 tumors. The most commonly gained segment was 8q22 approximately qter, found in all cases with 8q overrepresentation. Loss of parts of or the entire 8p was seen in 10 tumors. The smallest overlapping deleted region was localized to the subtelomeric region of 8p.

Hibernomas are benign tumors of brown fat, frequently characterized by aberrations of chromosome band 11q13. In this study, the chromosome 11 changes in five hibernomas were analyzed in detail by metaphase fluorescence in situ hybridization. In all cases, complex rearrangements leading to loss of chromosome 11 material were found. Deletions were present not only in those chromosomes that were shown to be rearranged by G-banding, but in four cases also in the ostensibly normal homologues, resulting in homozygous loss of several loci. Among these, the gene for multiple endocrine neoplasia type I (MEN1) was most frequently deleted. In addition to the MEN1 deletions, heterozygous loss of a second region, approximately 3 Mb distal to MEN1, was found in all five cases, adding to previous evidence for a second tumor suppressor locus in 11q13.

Pericentromeric rearrangements, such as isochromosomes and whole-arm translocations, are frequently encountered in short-term cultures from head and neck squamous cell carcinomas (HNSCC). To characterise further the localisation of the breakpoints in such rearrangements, metaphase cells from seven HNSCC known to carry structural rearrangements of the pericentromeric region of chromosome 5 were investigated using fluorescent in situ hybridisation (FISH) techniques. With a whole chromosome painting probe it could be confirmed that all chromosome 5 rearrangements identified at cytogenetic analysis contained chromosome 5 material. By using a centromere-specific alpha satellite probe it could be shown, however, that cytogenetically identical derivative chromosomes had different breakpoints. Thus, we conclude that the results of the present investigation add further support to the hypothesis that the essential outcome of near-centromeric chromosome rearrangements is the creation of genomic imbalances, i.e. gain and/or loss of neoplasia-associated genes.

The authors have developed a new method for constructing cosmid linking libraries. The method is based on the insertion of a selection gene, {beta}-lactamase, in genomic cosmid clones containing recognition sites for rare-cutting enzymes. The selection gene is maintained as a gene cassette in a plasmid and may be excised by the enzymes NotI, SacII, SplI, MluI, BssHII, and NarI or combinations of these enzymes. Using this gene cassette and a genomic cosmid library made from a human-hamster cell line containing the human chromosome 12 as its only human component, a chromosome 12-specific NotI linking library was constructed. The NotI linking clones contained recognition sites for other rare-cutting enzymes, SacII and BssHII, at high frequency, indicating the presence of CpG islands. Thirty cosmid linking clones were regionally localized by FISH and were found to be clustered to chromosome bands 12p13, 12q13, and 12q24. 28 refs., 2 figs., 1 tab.

The chromosomal integration of transfected circular plasmid pSV2neo molecules was investigated in normal and established mammalian fibroblasts, including two secondary cultures of normal mouse fibroblast lines, one established mouse-hamster hybrid cell line, R44, and the human fibrosarcoma line, HT1080. The physical organization of the integrated molecules was studied by restriction analysis. The results showed that whereas the normal fibroblasts predominantly integrated one head-to-tail partial dimeric molecule, the established cells predominantly integrated distinctly different molecular forms including deleted monomeres (HT1080) and various complex concatemeric molecules (R44), and frequently at more than one chromosomal site (R44). We also constructed a head-to-tail dimeric version of the plasmid, which in the case of the normal fibroblasts again integrated as a partial dimeric molecule in at least 50% of these cells. This result excluded the possibility that the normal mouse transfectants were selected for the integration of two functional neo genes. Thus, it is concluded that the distinctly different molecular forms integrated in normal and established cells demonstrate the operation of different integration pathways, the possible nature of which is discussed.

Cyclin D2, encoded by CCND2 at 12p13, takes part in the regulation of the cell cycle and has been suggested as a candidate for gene amplification in lymphoid malignancies. CCND2 is often overexpressed in chronic B-cell disorders, and we recently detected genomic amplification of the chromosomal region containing CCND2 in two of three investigated non-Hodgkin lymphomas (NHLs) with cytogenetic abnormalities involving 12p. In the present study, 58 NHLs without karyotypic evidence of 12p aberrations were analyzed by fluorescence in situ hybridization with probes for CCND2. No genomic amplification was found, strongly suggesting that this abnormality is rare in such NHLs.