Human Genetics

, Volume 133, Issue 4, pp 403–416

Recollections of a scientific journey published in human genetics: from chromosome territories to interphase cytogenetics and comparative genome hybridization


    • LMU Biozentrum
  • Christoph Cremer
    • Institute of Molecular Biology (IMB)
  • Peter Lichter
    • Division of Molecular GeneticsGerman Cancer Research Center (DKFZ)
Review Paper

DOI: 10.1007/s00439-014-1425-5

Cite this article as:
Cremer, T., Cremer, C. & Lichter, P. Hum Genet (2014) 133: 403. doi:10.1007/s00439-014-1425-5


In line with the intentions of an issue celebrating the 50th anniversary of Human Genetics, we focus on a series of frequently cited studies published in this journal during the 1980s and 1990s. These studies have contributed to the rise of molecular cytogenetics. They yielded evidence that chromosomes occupy distinct territories in the mammalian cell nucleus, first obtained with laser-UV-microbeam experiments and thereafter with chromosome painting, and contributed to the development of interphase cytogenetics and comparative genome hybridization. We provide a personal account of experimental concepts, which were developed by us and others, and describe some of the unforeseeable turns and obstacles, which we had to overcome on the way towards an experimental realization. We conclude with a perspective on current developments and goals of molecular cytogenetics.


The following recollections were written upon an editorial invitation to contribute to the 50th anniversary issue of Human Genetics for the reason that several of our publications in this journal during the 1980s and 1990s received a high number of citations. Frequent references, soon after a publication appears in print, are generally considered as an important indicator of its scientific impact, but this may be misleading and result in the neglection of important work. In line with editorial intentions of this anniversary issue, our recollections deal with studies of higher-order organization of chromosomes in the nucleus of mammalian cells, the development of interphase cytogenetics and comparative genome hybridization (CGH) published in Human Genetics. We wish, however, to deviate on purpose from a typical scientific review, written with the only intention to summarize the results of a research field during a certain period. Our intention is to present a personal account of the development of new concepts in the field of molecular cytogenetics and the efforts of experimental realization with their unexpected turns and obstacles. It appears that certain conditions need to be fulfilled before a new experimental field starts to flourish––the time must be ripe for it. The rise of molecular cytogenetics is no exception (for review see van der Ploeg 2000). Its basic methods were conceptually invented and realized by scientists working in Europe and the USA and our own story forms just a part of it.

Chromosome territories and their arrangements studied by laser-UV-microbeam experiments

Chromosome territories are an old and now generally accepted concept. It was first proposed in the late nineteenth and early twentieth century by Carl Rabl and Theodor Boveri but abandoned during the 1950s (for review see Cremer and Cremer 2006a, b), when early electron microscopic studies failed to distinguish chromosome territories in ultra-thin nuclear sections. At this time, the view became popular that only constitutive heterochromatin remains tightly condensed during interphase, whereas euchromatin strongly decondenses and intermingles. In the vein––or should one rather say vanity––of the rise of molecular biology a true understanding of important problems in cell and developmental biology was only expected from studies at the molecular level. The organization of genomes within the cell nucleus surely involved some packaging problems, but the nucleus was widely considered as a bag of intermingling chromatin fibers. Only few scientists argued for a dynamic nuclear architecture with fundamental functional implications for a cell type specific usage of genomes in normal and pathological cell types (Comings 1968; Vogel and Schroeder 1974; Blobel 1985). Our (TC’s and CC’s) first thoughts on experimental studies of chromosome arrangements in mammalian cell nuclei date back to 1968, when we were students of medicine (TC) and physics (CC) and discussed the potential of lasers as a research tool in cell and developmental biology. We wrote a pamphlet about our intention to build a UV-laser-microbeam and described a series of possible applications. In particular, we argued (too boldly as it turned out) that it should be possible with the help of UV-microbeam experiments to map the position of genes on chromosomes. For this purpose the microbeam should be directed to a chromosome of a living mitotic cell with the hope to generate cells with a distinct aberration on all descendants of the microirradiated chromosome. Although this hope never materialized, it prompted the development of micro-disk electrophoresis to identify variants of enzymes down to the single cell level (Cremer et al. 1972; Dames and Neuhoff 1972). Most important for the long-term direction of our research efforts, we argued that microirradiation experiments of interphase nuclei could help to elucidate higher-order chromatin arrangements: “Does an ordered arrangement of interphase chromosomes exist, which changes regularly with different states of function? Could it be that genes, which functionally belong together, are also positioned closely together, even if they are located on different chromosomes? If so, a method, which allows the isolated damage of small nuclear areas, might help to establish statistically significant “functional” linkage groups, which need not always be consistent with gene maps established by genetic experiments”. In early 1969, we distributed our pamphlet to a number of senior scientists, including Helmut Baitsch, Winfried Krone and Ulrich Wolf, in the hope to raise interest in our ideas. This was the case and in 1972 led to the formation of a small research team funded by the Deutsche Forschungsgemeinschaft (DFG) at the Institute of Anthropology and Human Genetics of the University of Freiburg i. Br., founded in 1961 and directed by Helmut Baitsch until 1970 and thereafter by Ulrich Wolf from 1972 to 2001. It is very rare that young researchers are allowed to start their own project, even before they have obtained a doctorate. This first ‘success’ came at the cost that we lacked any training in established methods for our new jobs and had to endure years with very few publishable results suffering from the permanent danger of complete failure. To our good luck our little team was greatly strengthened in the early by Christian Zorn, another young physicist and Jürgen Zimmer, a technical assistant with outstanding capabilities. Together we worked on the realization of the laser-UV-microbeam and on the development of experimental approaches to make damaged chromatin hit by the microbeam at a selected site of cell nucleus during interphase, visible in the subsequent mitosis. Such studies, of course, also required advanced 3D microscopy. CC started to think about possibilities of holographic focusing with the intention to overcome the limitations of the optical resolution (about 200 nm) of conventional light microscopy. In a manuscript published as part of a patent application we pointed out the principle possibility to build a Ho-Fo microscope (Ho-Fo stands for holographic focusing) and generate 3D images from cells with a resolution beyond the capacities of conventional light microscopy, based on the point-by-point recording of the fluorescence induced from the ultra-small focus (Cremer and Cremer 1972). In 1978, we proposed the construction of a laser scanning microscope for 3D imaging with conventional light optical resolution, where we added the idea of a Ho-Fo microscope in the appendix (Cremer and Cremer 1978). At this time, we lacked, however, the means and technical conditions to further pursue such goals. The first confocal laser scanning microscopes became available in the 1980s, whereas the experimental realization of super resolution fluorescence microscopy started during the 1990s (for reviews see Cremer and Masters 2013; Renz 2013).

In our 1969 pamphlet, we also considered microbeam experiments performed with egg cells and embryos. With the input of Klaus Sander and Christiane Nüsslein-Volhard, who worked then with Klaus Sander at the Institute of Zoology of the University of Freiburg, and Margit Lohs-Schardin, then a dedicated PhD student, the laser UV-microbeam apparatus was applied for localized irradiation of larval states and early Drosophila embryos resulting in fate maps based on location and frequency of defects seen in blastoderm stages and adult flies (Nusslein-Volhard et al. 1980; Lohs-Schardin et al. 1979a, b).

In 1976, we published our first paper on interphase chromosome arrangements in Human Genetics (Zorn et al. 1976). We microirradiated small nuclear regions (about 5 % of the total nuclear area) from diploid, fibroblastoid Chinese hamster cells, followed the cells from interphase to mitosis, prepared metaphase chromosome spreads in situ, and analyzed them for chromosome aberrations. To our disappointment, we rarely found chromosome aberrations in spite of the high energy density of the microbeam at its focal site. It was, however, possible to increase the yield of metaphase spreads with aberrant chromosomes dramatically, when we incubated microirradiated cells with caffeine (1–2 mM), which interferes with post-replication repair. Some metaphase spreads showed a few severely damaged chromosomes beside a majority of intact chromosomes. The damaged chromosomes lay typically together and were often affected to an extent that we found it appropriate to call them “shattered”. Accordingly, we referred to such events as partial chromosome shattering (PCS) and assumed that the shattered chromosomes were located in the nuclear area hit by the microbeam (Cremer et al. 1982a). In other mitotic cells, however, the entire chromosome complement appeared shattered or even pulverized (generalized chromosome shattering, GCS). Later studies supported the conclusion that mitotic cells with GCS represent a mitotic catastrophe resulting from a failure of normal chromosome condensation of the whole chromosome complement, including both microirradiated and non-irradiated chromosomes of a given nucleus (Hubner et al. 2009; Cremer and Cremer 1986, 2006b). Further microbeam experiments showed that the fraction of GCS increased with total UV-energy applied to a given nucleus independent of the energy distribution. For this purpose, we focused the beam either directly into the nucleus or above the nucleus such that the cone of the UV-microbeam irradiated the major part of the nucleus with the same incident energy (Cremer and Cremer 1986). To explain this phenomenon, we proposed a factor depletion model of generalized chromosome shattering (GCS) (see Fig. 14 in Cremer and Cremer 2006b). According to this model an unknown factor F, roaming the nuclear space, is required both at sites of post-replication repair (PRR sites) of DNA-photolesions and other sites generated in all chromosomes during S-phase. Competition of PRR sites for F can lead to a critical shortage of F at these other sites resulting in an inability of normal chromosome condensation when cells enter mitosis. Although chromosome condensation and segregation failed in such cells, they were able to form aberrant nuclei, but finally became eliminated by apoptosis. This little story exemplifies a mishap, which happens quite often, when a scientific journey is planned with a certain direction in mind. The original goal cannot be reached because the biological system reacts in an unexpected way. In our case, the experiments led to the discovery of a potential mechanism of a mitotic catastrophe, which can help to eliminate damaged cells.

To further pursue our journey in the intended direction, we developed an alternative approach. We microirradiated nuclei of Chinese hamster fibroblasts in G1 and incubated them, thereafter, for 2 h in medium with tritium labeled thymidine, which was incorporated into the damaged DNA by excision repair. Autoradiographs of cells fixed immediately, thereafter, showed clusters of silver grains over the microirradiated nuclear area (Zorn et al. 1979). We also microirradiated two nuclear areas remote from each other in G1. Following the incubation period with tritium thymidine, we allowed cells to proceed further in the cell cycle before fixation and autoradiography was performed. The relative positions of labeled chromatin were stably maintained. This result argued against large-scale changes of chromatin arrangements. Clusters of silver grains representing the microirradiated chromatin, however, became somewhat more dispersed (Fig. 1a). This result reflected in part an increase in nuclear size during interphase progression but also suggested constrained, small-scale movements of chromatin. This conclusion was confirmed by later live cell studies (Bornfleth et al. 1999; Walter et al. 2003; Strickfaden et al. 2010). We also allowed cells after microirradiation and labeling of a single nuclear area in G1 to proceed to mitosis. Autoradiographs of metaphase spreads, prepared in situ, revealed silver grains concentrated on a few chromosomes (Cremer et al. 1982b; Zorn et al. 1979) (Fig. 1b). Corresponding results were obtained, when we visualized microirradiated DNA of cells fixed immediately after microirradiation or at the next mitosis with a primary antibody specific for UV-irradiated DNA (Cremer et al. 1980; Hens et al. 1983). This approach made it possible to follow microirradiated chromatin both from interphase to the next mitosis and from mitosis to the next interphase (Cremer et al. 1984a). We were also able to demonstrate that the severely damaged chromosomes in metaphase spreads with PCS described above indeed represented the microirradiated chromosomes. In metaphase spreads with GCS, immunostaining was also restricted to a small area (Cremer et al. 1983). We could not detect any difference between shattered or pulverized chromosomes, which carried UV-induced DNA lesions and those, which did not. This finding supported our conclusion that GCS involved all chromatin of the microirradiated nucleus, including chromatin hit directly by the microbeam and chromatin located remote from the microirradiated nuclear area.
Fig. 1

Territorial organization of chromosomes demonstrated by laser-UV-microbeam experiments. a, b Autoradiographs of two Chinese hamster cell nuclei (adapted from Cremer et al. 1982b). Both nuclear poles were UV-microirradiated during G1. To achieve radioactive labeling of microirradiated DNA during excision repair, the cells were incubated with 3H-thymidine for 2 h. The nucleus in a represents an experiment, where fixation was performed immediately after this labeling period. Unscheduled DNA synthesis restricted to the microirradiated nuclear poles is indicated by compact clusters of silver grains (marked by arrows). The nucleus in b is representative for an experiment, where cells were allowed to grow for another 30 h in medium without 3H-thymidine before fixation. Clusters of silver grains are still restricted both nuclear poles. This finding argues against large-scale chromatin movements during the post-incubation period. Silver grains, however, have become somewhat more dispersed, indicating small-scale movements of chromatin. For further details and a quantitative evaluation see (Cremer et al. 1982b). Bars: 10 μm. c, d Autoradiographs of two metaphase spreads prepared in situ about 40 h following microirradiation in G1 at the nuclear edge and pulse-labeling with 3H thymidine for 2 h (Cremer et al. 1982b). In line with a territorial organization of chromosomes in the interphase nucleus, the autoradiographs demonstrate a few metaphase chromosomes intensely marked with silver grains. Silver grains are enriched over parts of these chromosomes, which were apparently hit by the microbeam, while other parts remained unlabeled, indicating that these parts were located outside the microirradiated nuclear area. Homologs of labeled chromosomes were typically unlabeled, indicating that the respective chromosomes territories were separated from each other. An interphase nucleus (a, left side) shows a cluster of silver grains at one nuclear pole. Bars: 10 μm

A study published 1982 in Human Genetics (Cremer et al. 1982b) was designed to distinguish between a territorial and non-territorial organization of interphase chromosomes in nuclei of cultivated Chinese hamster fibroblasts. We compared the higher-order chromatin arrangement in the interphase nucleus with a coil composed of individual threads (chromosomes). These threads may expand throughout the coil and intermingle extensively with each other (case A) or each thread may occupy its own distinct territory (case B). The two cases cannot be distinguished, when all threads bear the same, say white color, but a distinction becomes possible, when a small part of the coil is marked by red color before the coil is disassembled into individual threads (like chromosomes in a metaphase spread). In case A, short, red marks are scattered over many or even all threads, emphasizing their non-territorial organization in the coil (compare Fig. 5 in (Cremer and Cremer 2006b) and box 1 in (Meaburn and Misteli 2007). The territorial organization in case B is indicated by extended, red-colored segments limited to a few threads only, whereas the majority remains unmarked. The results of our experiments fully supported a territorial organization of interphase chromosomes in line with Carl Rabl’s and Theodor Boveri’s early prediction (Fig. 1c, d).

Chromosome territories and their arrangements studied by DNA–DNA in situ hybridization experiments

The microbeam experiments described above required special equipment, were tricky to perform and surely not useful for diagnostic applications. A real breakthrough for broad studies of chromosome organization and nuclear architecture with a wide range of diagnostic applications (see below) was achieved by the introduction of in situ hybridization techniques first with radioactive-labeled probes (Gall and Pardue 1969; John et al. 1969; Pardue and Gall 1969) and a few years later with chemically modified probes (Bauman et al. 1980; Rudkin and Stollar 1977). More and more human DNA probes were cloned. The introduction of biotin-labeled probes was a milestone in this development (Langer et al. 1981) and allowed the first studies of higher-order chromatin organization based on non-radioactive in situ hybridization experiments (Manuelidis et al. 1982; Manuelidis and Ward 1984).

Thrilled by these developments, one of the authors (TC) established isotopic in situ hybridization in his laboratory for studies of interphase chromosome arrangements. Probes specific for repetitive sequences located in the pericentromeric region of the human X and the long arm of the human Y provided the possibility to study the nuclear positions of these chromosomes in human lymphocyte nuclei. In nuclei carrying a t(X;Y) the two probes were found close to each other in line with the expectation that the translocation chromosome formed its own territory (Rappold et al. 1984). TC also had the good luck of getting training in non-isotopic in situ hybridization in the laboratory of Mels van der Ploeg in Leiden and during an EMBO course given 1984 in Paris by Laura Manuelidis and Giorgio Bernardi. Chinese hamster x man and mouse x man hybrid cells provided model systems, which allowed for the first time the direct visualization of human chromosome territories by radioactive and non-radioactive in situ hybridization using total human genomic DNA as a probe (Manuelidis 1985; Schardin et al. 1985). The rationale of this approach was based on evidence that these mammalian species contain a large amount of evolutionary divergent, repetitive sequences, which do not cross-hybridize on chromosomes of other species.

How much evidence supporting a functionally important topography and organization of the (re-) discovered chromosome territories was achieved by all these experimental efforts? Embarrassingly little, as Widukind Lenz, one of Germany’s leading human geneticists, pointed out in a letter, which he wrote to TC in 1986 after listening to a talk on chromosome topography. “When you write “topography” you may all too easily create the impression that a disparity exists between a pretension and the actual experimental proof. Although I am not personally competent in this field, I agree with the opinion of others, who understand more about this problem than I do. It seems to me that evidence for a non-random distribution of chromosomes in the nucleus does not yet argue for a topography. Differences in size, replication and base pair sequences may in many respects result in deviations from a random distribution, which should not be referred to as a topography, but rather reflect circumstances, which are just not identical for all chromosomes. It further appears to me that the mirror-like position of homolog chromosomes (in daughter nuclei) briefly after mitosis may not provide any hint to a generally applicable topography. I feel that you have already spent a lot of time and acumen on this problem, which you might have used elsewhere in more fruitful ways. I readily admit that great discoveries are not all that rare made by attacking problems, which were considered as unpromising or futile. Unfortunately, the experts may be right after all in most cases”.

The letter’s addressee did not follow this well-intended advice, in part because of some deeply ingrained stubbornness to follow one’s own route towards hoped-for discoveries, in part because it was not easy to break off a scientific journey after an investment of some 15 years, but in particular because of encouraging discussions with Laura Manuelidis on the unsolved problem of nuclear architecture in Paris, which resulted in an invitation to come to Yale and work with her and David C. Ward on the problem. Our discussions were reflected in a successful application of TC for a Heisenberg stipend, a highly esteemed fellowship given by the DFG for 5 years to promising, young researchers—in this case an already 40-year-old one trying to maintain hope for a tenured scientific career. The application described the current state and perspectives of chromosome topography, a plan to establish a FISH procedure for the specific visualization of any human chromosome or chromosomal subregions of interest and the importance of such a procedure for experimental and clinical cytogenetics. At the same time TC and PL discussed possibilities to realize these plans. A successful application of PL for an EMBO fellowship made it possible for both of us to come to Yale in 1986 and join forces with Laura and David.

Chromosome painting

The plan to visualize entire, individual human chromosomes both during mitosis and interphase, an approach later appropriately called chromosome painting (see below), was based on the availability of DNA libraries from flow-sorted human chromosomes. One of us (CC) had contributed at the Lawrence Livermore National Laboratories to preparative dual-beam sorting of the human Y chromosomes, which allowed the generation of a human Y-specific DNA library (Cremer et al. 1982c; Muller et al. 1983; Cremer et al. 1984b). In 1986, Dan Pinkel together with his colleagues Tore Straume and Joe W. Gray published a landmark paper on “Cytogenetic analysis using quantitative, high-sensitivity, fluorescence hybridization” (Pinkel et al. 1986). Since then fluorescence in situ hybridization (FISH) has become the method of choice for non-isotopic in situ hybridization, because of its potential for multicolor FISH (see below).

These methodological breakthroughs set the stage for the realization of the research plans, which we (TC and PL) had developed together with Laura Manuelidis and David C. Ward. We wanted to use then available libraries to generate complex probes for visualization of the respective chromosomes in FISH experiments but had to overcome a major obstacle to do so; each library also contained repetitive sequences dispersed through the entire chromosome complement. We considered two possibilities to generate chromosome specific “paint” probes from the DNA libraries established from sorted chromosomes. One approach was based on the physical removal of dispersed, repetitive sequences from such libraries, the other approach on the use of an excess of unlabeled carrier DNA designed to prevent the hybridization of labeled, repetitive sequences to non-target chromosomes, such as DNA from hybrid cells carrying many human chromosomes except the chromosome selected as the target of “painting” or the fast reassociating fractions of whole genomic DNA carrying highly repetitive human DNA sequences. Whereas the affinity-based removal of repetitive DNA failed at this time despite great efforts, chromosome painting was successfully established with hybridization mixtures using library DNA from sorted chromosomes plus an excess of unlabeled Cot-1 DNA or even total human genomic DNA (Lichter et al. 1988a; Cremer et al. 1988) (Fig. 2). Successful protocols to remove unwanted repetitive sequences from chromosome painting probes became much easier after the advent of PCR and were established several years later (Craig et al. 1997). The fitting term “chromosome painting” was invented by Joe Gray and Dan Pinkel, who together with other members of their laboratories pursued the visualization of individual chromosomes based on the same concept (Pinkel et al. 1986). The fact that different groups developed basically the same concepts and pursued them independently underlines that the time was ripe for this development. On a personal note we wish to emphasize the openness of intellectual exchanges on current progress, when we met on conferences and the fact that Dan Pinkel was the first to show a lovely image of a painted chromosome in a non-hybrid cell.
Fig. 2

Chromosome territories demonstrated by chromosome painting. Chromosome painting was carried out with biotinylated DNA from sorted human chromosomes 7. Hybridization of biotinylated, repetitive sequences to repetitive targets dispersed throughout the entire human genome was suppressed with an appropriate amount of unlabeled, human genomic competitor DNA added to the hybridization mixture. For details see Lichter et al. (1988a). Bar: 10 μm. Metaphase spread a and interphase nucleus b with two painted chromosomes 7 (Lichter et al. 1988a). Metaphase spread c and interphase nucleus d from a human glioblastoma cell line show five painted HSA 7 chromosomes (Cremer et al. 1988). The arrow in c points to an aberrant chromosome 7, which could be identified as an iso(7p) (for details see Cremer et al. 1988)

A turn to interphase cytogenetics

In a letter written to David C. Ward during summer 1985 TC wrote: “Of course, together with Laura (Manuelidis) unravelling the problem of interphase topography has been and will remain my particular aim. However, I feel it intriguing that the procedures, which have to be developed for this particular aim at the same time appear to have many other applications as well. … The more I think about possible applications of a procedure to selectively stain specific chromosomes or chromosomal subregions in the interphase nucleus the more I become convinced that such procedures will become of utmost importance not only for studies of the interphase chromosome topography but also for many applications in clinical cytogenetics ranging from prenatal diagnosis to tumor cytogenetics… My thinking is based on the evidence for a territorial organization of chromosomes in the nucleus of a cell. For example, we have recently shown (our unpublished data) that it is possible to demonstrate three chromosomes 18 in cells from a fetus with trisomy 18 by using a probe specific for the constitutive heterochromatin (pericentric DNA) of this chromosome. Given enough cloned DNA probes and pools of probes, respectively, which hybridize to specific chromosomes, trisomies and monosomies of these chromosomes, as well as duplications or deletions of subregions or even specific translocations could be identified directly in the interphase nucleus. … Considering the clinical significance of chromosomal changes, for example in acute lymphoblastic (ALL) and non-lymphoblastic leukemias (ANLL), where specific translocations are not only of great importance for tumor classification but also of great prognostic value (including the duration of remission after initial therapy and the mean survival time) I am convinced that detection (with automated procedures) of such specific abnormalities directly in the interphase nucleus from peripheral blood cells or bone marrow cells might greatly facilitate diagnosis and supervision of patients during therapy and remission. In contrast to the classic evaluation procedures using banded chromosomes, the new procedure would hopefully allow the detection of even very small cell clones with specific chromosomal aberrations. There are many other examples one can think about, such as cytogenetics of cells from solid tumors, where it might be very difficult to obtain any well spread metaphases. Staining of specific chromosomes by pools of chromosome specific sequences might be used to test whether a certain chromosome has participated in the formation of unidentified marker chromosomes etc. While I am convinced, of course, that classic cytogenetics will have its important applications in any foreseeable future, I am also convinced that cytogenetics performed directly on the cell nucleus of interphase cells or even postmitotic cells will have many broad applications in the future and open highly interesting new avenues for experimental and clinical cytogenetics” (Fig. 3a, b).
Fig. 3

Concept of interphase cytogenetics. Schemes were prepared by T. Cremer in 1985 to point out the potential usefulness of bi-colored in situ hybridization assays for the detection of chromosome aberrations in both metaphase spreads and interphase nuclei. a, b Chromosomal subregions of interest are marked in different colors by in situ hybridization with “nested sets of chromosome specific DNA probes” (Cremer et al. 1986): “Normal interphase nuclei should present two green and two red fluorescent spots with variable distances. Consider a specific translocation between these chromosomes with breakpoints in the defined subregion of the chromosomes. We predict that interphase nuclei containing the specific translocation should bear three green and three red fluorescent spots. The distribution of each set of spots is expected to be variable depending on the presumably variable positions of the chromosomes involved. However, two green and two red spots should be detected side by side, since they mark the positions of the two translocation chromosomes. The third green and red fluorescent spot, respectively, would indicate the position of each normal homologue”. c, d The differential coloring of entire chromosomes (chromosome painting) should allow metaphase and interphase detection of rearrangements between these chromosomes independent of the site of the breakpoints

In 1986, joint efforts of the groups of T. Cremer, Mels van der Ploeg and Peter Pearson led to a publication in Human Genetics on interphase cytogenetics (Cremer et al. 1986). We demonstrated the detection of a trisomy 18 in amniotic fluid cell nuclei and discussed the possibilities of an interphase diagnosis of chromosomal translocations by in situ hybridization with breakpoint specific probes (see legend to Fig. 3a). The new possibility of interphase detection of the most common, numerical chromosome aberrations in prenatal diagnosis, the trisomies 13, 18 and 21 (Lichter et al. 1988b; Boyle et al. 1990) raised hopes to circumvent more invasive diagnostic procedures by diagnosing these trisomies in the small portion of fetal cells within the peripheral blood of pregnant woman. Despite many efforts by multiple laboratories, a routine application of this concept failed, however, due to difficulties in the enrichment of the fetal cells. Chromosome painting with or without the additional use of subregional probes was soon established as an important add-on to chromosome banding (Jauch et al. 1990; Lengauer et al. 1992; Popp et al. 1993). Among the important diagnostic applications was a FISH test developed in Marion Cremer’s laboratory to prove or rule out a carrier status of mothers, who had given birth to a boy with Duchenne muscular dystrophy (Ried et al. 1990; Tocharoentanaphol et al. 1994) (Fig. 4). This test attracted many affected families both from Germany and abroad. If a mother turned out to be a proven carrier, other female members of these families were tested as well.
Fig. 4

A FISH assay for the detection of female carriers of Duchenne muscular dystrophy. a Two-color FISH of a diploid human fibroblast nucleus (46, XX) with a human X-specific paint probe and a cosmid probe specific for the dystrophin gene located at Xp21. Both X-chromosome territories (red) show the green dystrophin gene signal (arrows). Bar: 10 μm. b) Two red painted X-chromosomes from a female carrier. In addition to the X-specific paint probe a cosmid probe specific for the region deleted in her son suffering from Duchenne muscular dystrophy was used. Another cosmid probe, which maps to Xq28, was employed as an internal control for hybridization efficiency. The left X shows hybridization signals on Xp21 and Xq28. On the right X only the control signal is visible. This pattern was consistently observed in numerous metaphase spreads proving the carrier state of the mother and excluding a de novo deletion in the affected boy. (adapted from Cremer and Cremer 2006b; example provided by Marion Cremer and Anna Jauch). For details, see Ried et al. (1995) and Tocharoentanaphol et al. (1994). Bar: 10 μm

Comparative genome hybridization: a tool to study gains and losses of genetic material

FISH is clearly not suitable to screen cells for unknown chromosome aberrations without a hint where to look for. Before one can start a FISH experiment one has to make a proper selection of probes useful to detect specific aberrations. The development of comparative genome hybridization (CGH) has allowed genome-wide studies of chromosomal imbalances at the level of chromosomes and chromosomal subregions. The first publication on this approach appeared in Science in 1992 (Kallioniemi et al. 1992) followed a few months later by publication in Human Genetics from the T. Cremer and Lichter groups (du Manoir et al. 1993). The principle of CGH is based on the hybridization of equal amounts of differentially labeled whole genomic test and control DNA probes to metaphase spreads with a normal karyotype. Fluorescence ratios of the hybridization signals of the test DNA, e.g., DNA prepared from a tumor sample, and control DNA, e.g., DNA prepared from diploid cells, are measured along chromosomes. For chromosomes present in two copy numbers in the test cell sample one expects a fluorescence ratio normalized to 1. Significant increases above 1 along an entire chromosome or over a certain segment indicate a copy number increase of the entire chromosome or the respective segment, while significant decreases delineate a specific loss. The method turned out to be highly suitable for genome-wide tests of chromosomal imbalances and allowed such studies even using amplified DNA from formalin-fixed tumor tissues (Ried et al. 1995; Speicher et al. 1993, 1995). The resolution limit of CGH was in the order of 2–10 Mbp (Bentz et al. 1998; du Manoir et al. 1995). The development of array CGH, however, increased this resolution to the level of individual genes and even beyond (Pinkel et al. 1998; Solinas-Toldo et al. 1997). Recurrent genetic imbalances in tumor samples identified regions in the human genome where a search for oncogenes or tumor suppressor genes was indicated.

Considering the worldwide use of CGH and array CGH demonstrated by many thousands of publications, a reader of a historical account may tend to expect that their implementation was a straightforward event both conceptually and experimentally because the time was ‘ripe’ for it. But, in fact, this part of our journey took about a decade between the first conceptual thoughts and their realization. In 1983, one of us (TC) described some thoughts in his private laboratory journal under the heading “Development of a system for the analysis of genomes for chromosomal duplications (partial trisomies) and deficiencies (partial deletions)”. “Is it possible to visualize differences between (A) one copy of a single-copy sequence per genome, (B) two copies, (C) three copies? … Put DNA probes as spots on a filter and hybridize nick-translated genomic DNA from a patient. Probes can be spotted in a well arranged format. The number of probes can be increased nearly without limit. In case of 500 to 1,000 spots from different single-copy probes, which are uniformly distributed over the genome, one could reach a resolution similar to the resolution of a high quality chromosome banding analysis. This analysis could be used to localize duplications and deficiencies on chromosomes in cases where such identification fails with conventional microscopic techniques. It is not clear, whether the sensitivity of spot-hybridization can be increased to the point that differences in the blackness of spots can be unambiguously detected by autoradiography depending on the presence of the hybridized sequence in one, two or three copies per genome. If such a sensitivity can be reached, one should expect that depending on the size of a duplication or deletion, neighboring spots on a special filter should also be stronger or weaker than control spots representing non-affected regions of the genome… Summary: In near future it should become possible to use cloned, chromosomally mapped, single-copy sequences for the detailed analysis of breakpoints or duplication-deficiencies in cases, where the participating chromosomes can be identified with conventional banding techniques. In the distant future it may become possible to make a cytogenetic diagnosis in case of a genetic indication (clinical hints on a chromosomal imbalance without microscopically identifiable chromosomal changes) or in case of a microscopically identified aberration without the possibility to identify a small translocated fragment beyond the possibilities of classic microscopic techniques”. A large set of cloned, single-copy probes was not yet available at that time. Furthermore, the problems to develop a CGH method based on radioactivity, seemed too complex and most cytogeneticists were apparently quite happy with the potential of high-resolution chromosome banding. In short, the time for such a development was not ripe in 1983 and the idea was put to rest. In 1987, PL and TC reflected on ways and means to generate large amounts of the necessary single-copy probes. At this time, PCR had started to conquer all genetic laboratories worldwide. Accordingly, we argued that it should become possible to amplify DNA pieces from selected chromosomal regions simultaneously, provided that the necessary sequence information became available for the synthesis of the oligonucleotides necessary for this kind of multiple, simultaneous polymerase chain reaction. The amplified sequences should be put in a dot like format on an appropriate filter. We considered the possibility to start with the development of a test to search for deletions in the X-linked gene responsible for Duchenne muscular dystrophy and later extend this approach to the genome-wide analysis of duplications and deletions. We also considered the possibility to use such an approach for the identification of polymorphisms of amplified DNA fragments in human genomes. These ideas were laid down in an unpublished manuscript but again put to rest. The technical problems to develop such an approach to the point, where such applications would become feasible, required large investments and appeared still unsurmountable at least for us.

In a letter written in September 1990 to Lore Zech, TC considered a solution, which was technically less demanding. “When one has only DNA available from a solid tumor and normal good metaphase spreads from normal lymphocytes, is it possible to figure out, whether certain chromosome segments are underrepresented (e.g., deletions of chromosome arms) or overrepresented (e.g., trisomies)? The answer depends on the question, how precisely one can do a ratio-imaging with a CCD camera or laser scanning (microscopy) and digital image analysis. Experiment: direct-labeling of normal, human DNA, e.g., with FITC, direct labeling of tumor DNA, e.g., with AMCA, double (in situ) hybridization to normal metaphase spreads. … In case of a double hybridization experiment with tumor DNA and normal DNA to normal metaphase spreads one should then find a significantly increased ratio of AMCA/FITC over chromosome segments, which are overrepresented in the tumor compared with non-affected chromosomes. Over segments, which are underrepresented in the tumor, the ratio should be significantly lower…. For suppression I use unlabeled Cot-1, which is commercially available and works very well. The trick is to make relative measurements against the internal standard. The differences to be expected are, of course, small. For trisomic regions xxx-AMCA/xx-FITC = 1.5, for control regions xx-AMCA/xx-FITC = 1, for deleted regions x-AMCA/xx-FITC = 0.5. I do not think that one can recognize such differences with the unaided eye. But it may be possible to recognize them with ratio-imaging.” At this time the TC and PL laboratories joined forces. The concept was realized in 1992 and published in Human Genetics in early 1993.

Scientific journeys cannot be undertaken without funding. In retrospect a reader may tend to believe that the two publications on CGH from Kallioniemi et al. in October 1992, and our own publication in February 1993 should have immediately opened the door for getting funding for further work. But this was not the case. In fact, work on CGH was carried out in our laboratories for several years without dedicated funding. Every scientist, who pursues long-term projects, knows that rejections of his or her grant applications are an indispensable part of a scientific career. We mention this aspect not because of any feelings of being not well treated personally. We received fellowships and grants at critical periods of our scientific careers and we recognize this good luck most thankfully. We wish to mention this aspect in our historical account for a more general reason. In a retrospective view judgements about the long-term importance of new directions of research with new scientific methods seem to be easy, but this is not the situation of reviewers, who have to judge a grant application, when such a new development starts. Reviewers may not easily be convinced to support applications, which do not fit on the current band waggon in the figurative sense of goals and methods considered important by the scientific community at a given time. In early 1993 we (TC and PL) submitted a grant application “Detection of genetic imbalances in cancer cells: development of an automated test system based on comparative genomic in situ hybridization and digital image analysis” to the European Community, who had implemented a program Biomed I for biomedical and health research. We thought that we had brought together an impressive international collaboration of scientists with industrial partners, The application was written together with Michel Robert-Nicoud (Grenoble), and supported by further French and Italian scientists, as well as by the Carl Zeiss Company, Oberkochen, Germany, and by Alcatel, a French software company. But the application was rejected. In 1996, we made a second attempt to achieve funding from the European Community with an application this time within the realms of the Biomed II program: “High Resolution Molecular Diagnostic using Automated Matrix Comparative Genomic Hybridization“. Participants in addition to TC and PL were Aaron Bensimon (Pasteur Institute, Paris, France), Michael Speicher (TC laboratory), Jan M.N. Hoovers (Institute of Human Genetics, Academic Medical Centre, Amsterdam, The Netherlands), as well as Applied Imaging International, UK, as an industrial partner. Again the grant application was rejected. Nonetheless, this second application received a positive comment”: The research plans are realistic and feasible. However, the project is neither relevant for areas 2.2 (Research on imaging systems and other devices and techniques for diagnosis and therapy in order to increase diagnostic power and possibilities of simultaneous intervention) nor to area 5 (Human genome research) to which it (was) sent for application.” As a take home message for young researchers, we suggest not to wait for official permission of granting agencies, if you want to do something new. Go ahead, if you can do it as a side track of running grants.


Molecular cytogenetics has played an important role to pinpoint the location of genes on chromosomes (Landegent et al. 1987; Lichter et al. 1990). The localization of oncogenes, in particular, has resulted in major progress towards understanding the genetic origins of cancer. Methods, which allow the simultaneous metaphase and interphase visualization of all chromosomes in different colors (Bolzer et al. 2005; Schrock et al. 1996; Speicher et al. 1996) have greatly enhanced the possibilities of diagnostic bi-color detection schemes envisaged in 1985 (Fig. 3). Considering the enormous progress of high-throughput approaches to study genomes and transcriptomes of normal and pathological cells at ever decreasing prices (Hudson et al. 2010), one may ask, however, whether microscopy based cytogenetics will continue to play an indispensable role in future research and diagnostics. We are convinced that this will be the case. The profound cell-to-cell variability of nuclear architectures during development and cell differentiation provides a major reason for the need of single cell analyses. Although molecular biological approaches are being developed to study genomes and transcriptomes at the single cell level, microscopy has unique advantages. The 3D resolution, which has become possible with super resolution fluorescence microscopy (for reviews see Cremer and Masters 2013; Renz 2013), is currently approaching the low nanometer range with applications not only in fixed but also in living cells. This resolution combined with ultrasensitive methods for the simultaneous detection of macromolecules and macromolecular assemblies of interest may eventually allow a routine cytogenetic diagnosis of genetic changes down to point mutations and accompanying epigenetic changes in individual nuclei. Multicolor formats will allow the simultaneous detection of multiple DNA, RNA and protein markers which characterize genetically normal and aberrant cells located side-by-side within the natural landscape of tissues. In addition to the presence or absence of a combinatorial set of markers with diagnostic and prognostic value, we expect that their nuclear topography will add further important diagnostic information.

Studies of nuclear architecture have made great progress during the last decade, including the development of methods for the visualization of chromosome territories, chromatin domains and genes in living cells (Newhart and Janicki 2014; Strickfaden et al. 2010; Tsukamoto et al. 2000; Walter et al. 2003; Zink et al. 1998). Whereas the importance of epigenetic modifications for the cell type specific usage of genomes has been unequivocally demonstrated during the past decade, the contribution of changes of chromosome topography to changes of gene expression patterns during development and cell differentiation has remained an open question. The development of our own model views of the functional nuclear architecture during the last two decades has been described in a series of reviews (Cremer et al. 1993, 2000, 2006; Cremer and Cremer 2001, 2010; Lanctot et al. 2007). A comprehensive understanding of this problem requires a quantitative approach to explore the space–time dynamics of architectural changes during development and differentiation, including positional changes of endogenous loci and chromosome territories (CTs), as well as interactions between chromatin sites, nuclear bodies and nuclear machineries involved in transcription, co-transcriptional splicing, DNA/chromatin replication and repair. Only on such a fundament a systematic search can be initiated for molecular mechanisms, which connect established changes of nuclear architecture with changes of nuclear functions, such as changes of gene expression patterns.

Super resolution fluorescence microscopy has started to provide 3D images of nuclear architecture with unprecedented resolution (Schermelleh et al. 2008; Markaki et al. 2010, 2012). Multicolor staining of different nuclear components is a key to topographical studies aimed at an understanding of their dynamic nuclear arrangements in space and time. Because of its unsurpassed resolution, electron microscopy will remain an important approach towards understanding the structural biology of cell nuclei in a functional context (for review see Rouquette et al. 2010). Methods of correlative microscopy, which allow the sequential visualization of one and the same cell using a series of microscopic approaches with increasing resolution allow to take advantage of the particular strength of each approach and compensate for its specific limitations (Hubner et al. 2013). Current developments of molecular biological approaches (de Graaf and van Steensel 2013; Dixon et al. 2012; Gibcus and Dekker 2013; Cavalli and Misteli 2013; Bickmore and van Steensel 2013; Nagano et al. 2013) encourage a vision of a future, in which a combination of microscopic and molecular biological tool sets will allow high-resolution studies of nuclear landscapes in individual structurally intact, even living cells.

Finally, we wish to emphasize the importance of evolutionary comparisons to gain insight into basic structural principles of higher-order nuclear organization and their functional implications. Chromosome painting has allowed the reconstruction of chromosomal rearrangements during primate evolution (Jauch et al. 1992; Scherthan et al. 1994; Wienberg 2004) and also provided first evidence for an evolutionary conservation of higher-order chromatin arrangements despite profound chromosomal rearrangements (Tanabe et al. 2002). The new tool sets provide unprecedented possibilities to explore evolutionary conserved features of nuclear architectures throughout the whole range of eukaryotes and to compare them with nucleoids in prokaryotes as outgroups. Such studies will provide new insights into the problem of selective constraints that played a major role during the evolution of larger and larger genomes and finally provide a conclusive answer to the possible functional relevance of a 4D (space–time) genome organization.


Our scientific journey (and our careers) would have failed without the essential ideas, the experimental skills and the continuous support of colleagues and friends. The support of Winfried Krone, Ulrich Wolf and Helmut Baitsch was essential to form a little laser UV-microbeam research group in the early 1970s at the Institute of Anthropology and Human Genetics of the University of Freiburg i. Br.. TC and CC wish to thank, in particular, Christian Zorn and Jürgen Zimmer for their invaluable contributions to our early UV-microbeam studies of nuclear architecture. TC is indebted to his mentor Friedrich Vogel, director of the Institute of Human Genetics and Anthropology of the University of Heidelberg from 1962 to 1993, for his unfailing support in critical times, and to Traute Schroeder-Kurth, for the opportunity to establish his own laboratory within the Institute’s Divison of Cytogenetics directed by her. Opportunities to work as guest researchers in laboratories providing conceptual and experimental expertise in cutting-edge technologies mark turning points in our scientific endeavors. CC thanks Marvin A. van Dilla and Joe W. Gray for the opportunity to work with them at the Biomedical Sciences Division of the Lawrence Livermore National Laboratories on the sorting of the human Y chromosome. TC and PL greatly acknowledge support from Laura Manuelidis and David C. Ward. Chromosome painting and the direct demonstration of chromosome territories were achieved together with them in their laboratories at Yale University. In addition, TC thanks Michael W. Berns (University of California, Irvine), Wolfgang Hennig and Peter H. Vogt (then University of Nijmegen), Mels van der Ploeg (University of Leiden), Maynard V. Olson and Harold C. Riethman (then Washington University, St. Louis) for making it possible to learn and perform bench work in their laboratories. Interphase cytogenetics was introduced during the 1980s in TC’s Heidelberg laboratory with support from Peter Pearson, Mels van der Ploeg and their co-workers and essential, experimental contributions (in alphabetical order) from Marion Cremer, Patricia Emmerich, Anna Jauch, Christoph Lengauer, Susanne Popp, Gudrun Rappold and Margit Schardin. The application of chromosome painting as a tool to study chromosomal rearrangements during evolution was initiated by Johannes Wienberg. Daniela Zink performed the first studies of chromosome territories in living cells. The concept of CGH was realized and applied to tumor samples in the TC and PL laboratories with the help of numerous colleagues, including (again in alphabetical order) Stanislas du Manoir, Stefan Joos, Thomas Ried, Harry Scherthan, Evelin Schröck, Michael Speicher and Ruthhild Weber. A first realization of the concept of array CGH (or matrix-based CGH as we originally called this method) was achieved in the laboratory of PL with major impact from Sabina Solinas-Toldo and colleagues. Finally, we wish to thank the Deutsche Forschungsgemeinschaft (DFG), who supported our work with fellowships and grants. Walther Klofat represented for us the good spirit of the DFG in person. Excerpts cited from grant applications, letters and unpublished manuscripts written in German were translated in English from TC.

Copyright information

© Springer-Verlag Berlin Heidelberg 2014