Nuclear texture parameters as discriminant factors in cell cycle and drug sensitivity studies

The authors have recently shown that cell cycle characteristics of in situ cell populations can be determined using the SAMBA 200 cell image processor by computing 15 densitometric and texture parameters on each Feulgen-stained nucleus and multiparametric analysis of data. The present paper displays the importance of chromatin pattern assessment and detection of conformational changes in DNA structure, based on nine nuclear texture parameters measured from the grey level cooccurrence and the run-length section matrices. Reference files were constructed by merging respective reference files (G0/G1, S, G2 and M) of MDA AG and MCF-7, two mammary epithelial cell lines presenting different morphological aspects and hormone responses, these files were found to be valid in the reclassification of any mammary epithelial cell in culture with a diploid or near diploid pattern. Moreover, the authors demonstrate that chromatin texture changes, following direct interaction of chemotherapeutic drugs with DNA, may be assessed owing to nuclear texture parameters. Consecutive to daunomycin addition (0.5 microgram/ml) and concomitant to the appearance of nuclear morphological alterations in MDA AG sensitive cells as viewed by microscopic observation, discriminant factorial analysis showed progressively increasing erroneous reclassification from 15 to 72 h of treatment. These experimental results were exploited with a kinetic mathematical model to quantify the daunomycin blocking effect: 20% in S phase and 80% in G2 phase. Interestingly, no textural change was observed on MDA A1 anthracycline resistant cells, indicating that these texture parameters may permit distinction of drug sensitive cells. This methodology 1) can be applied to test in vitro resistance-reversal molecules, 2) may be extended to other therapeutic agents giving rise to conformational changes in DNA structure, and 3) can be applied to cytopunctions or imprints of tumor biopsies with diploid-like DNA content to follow evolution of drug sensitivity or resistance during course of therapy.     

Structural features of the first component of human complement, C1, as revealed by surface iodination.

Lactoperoxidase-catalysed surface iodination and sucrose-gradient ultracentrifugation were used to investigate the structure of human complement component C1. 1. Proenzymic subcomponents C1r and C1s associated to form a trimeric C1r2-C1s complex (7.6 S) in the presence of EDTA, and a tetrameric Clr2-C1s2 complex (9.1 S) in the presence of Ca2+. Iodination of the 9.1 S complex led to a predominant labelling of C1r (70%) over C1s (30%), essentially located in the b-chain moiety of C1r and in the a-chain moiety of C1s. 2. Reconstruction of proenzymic soluble C1 (15.2 S) from C1q, C1r and C1s was partially inhibited when C1s labelled in its monomeric form was used and almost abolished when iodinated C1r was used. Reconstruction of fully activated C1 was not possible, whereas hybrid C1q-C1r2-C1s2 complex was obtained. 3. Iodination of proenzymic or activated C1 bound to IgG-ovalbumin aggregates led to an equal distribution of the radioactivity between C1q and C1r2-C1s2. With regard to C1q, the label distribution between the three chains was similar whether C1 was in its proenzymic or activated form. Label distribution in the C1r2-C1s2 moiety of C1 was the same as that obtained for isolated C1r2-C1s2, and this was also true for the corresponding activated components. However, two different labelling patterns were found, corresponding to the proenzyme and the activated states.     

Characterization of a plasmin-digest fragment of rabbit immunoglobulin gamma that binds antigen and complement.

Rabbit immunoglobulin gamma (IgG) was digested with plasmin after being left for 15 min at pH2.5, 30 degrees C followed by a rapid increase in the pH to 7. The fragment antigen and complement binding (Facb) was isolated and characterized chemically and biologically. Sequence studies showed that the C-terminal quarter of the heavy chain had been removed, the split occurring at a lysine-alanine bond in the sequence Thr-Ile-Ser-Lys-Ala-Arg. The fragment Facb retained the capacity to precipitate with antigen and the precipitate caused activation of the first component of complement of the same order as that of acid-treated IgG. Both Facb and acid-treated IgG showed a fall in complement fixation relative to the native molecule of 30-40%.     

35S-Atractyloside binding affinity to the inner mitochondrial membrane

Isolated inner mitochondrial membrane contains a small number of binding sites for atractyloside (of the order of 0.1 nmole/mg of protein) with very high binding affinity (half saturation at 0.014 &mgr;M atractyloside). The high affinity binding ability of the inner mitochondrial membrane is markedly decreased upon aging, acidification of the medium or addition of ADP, but remains unchanged in the presence of uncouplers such as FCCP. Added ADP causes a two-step transition from the high affinity binding to low affinity binding (K(d) > 0.50 &mgr;M) and concomitantly a significant increase of the measured number of binding sites (about a doubling). The half maximum effect in the first step transition is given by 1 &mgr;M ADP. The use of 35S-atractyloside as a probe of the inner mitochondrial membrane conformation specifically related to the adenine nucleotide translocation is discussed.     

Antibody-independent interaction between the first component of human complement, C1, and the outer membrane of Escherichia coli D31 m4.

The heptoseless mutant of Escherichia coli, E. coli D31 m4, binds C1q and C1 at 0 degrees C and at low ionic strength (I0.07). Under these conditions, the maximum C1q binding averages 3.0 X 10(5) molecules per bacterium, with a Ka of 1.4 X 10(8) M-1. Binding involves the collagen-like region of C1q, as shown by the capacity of C1q pepsin-digest fragments to bind to E. coli D31 m4, and to compete with native C1q. Proenzyme and activated forms of C1 subcomponents C1r and C1s and their Ca2+-dependent association (C1r-C1s)2 do not bind to E. coli D31 m4. In contrast, the C1 complex binds very effectively, with an average fixation of 3.5 X 10(5) molecules per bacterium, and a Ka of 0.25 X 10(8) M-1, both comparable with the values obtained for C1q binding. C1 bound to E. coli D31 m4 undergoes rapid activation at 0 degrees C. The activation process is not affected by C1-inhibitor, and only slightly inhibited by p-nitrophenyl p'-guanidinobenzoate. No turnover of the (C1r-C1s)2 subunit is observed. Once activated, C1 is only partially dissociated by C1-inhibitor. Our observations are in favour of a strong association between C1 and the outer membrane of E. coli D31 m4, involving mainly the collagen-like moiety of C1.     

Characteristics of complement subcomponents C1r and C1s synthesized by Hep G2 cells.

The association and activation states of complement subcomponents C1r and C1s biosynthesized by Hep G2 cells were studied. C1r and C1s are secreted in stoichiometric amounts; in the presence of Ca2+ they are associated in a complex that sediments similarly to plasma C1r2-C1s2. Both compounds are synthesized as monomer proteins of apparent Mr 86 000. C1r is secreted as a dimer. Secreted C1r is not autoactivatable but undergoes proteolysis by exogenous C1r; secreted C1s is also proteolysed by exogenous C1r. In the presence of immune-complex-bound C1q, secreted C1r and C1s are able to reconstitute C1, but normal activation requires extrinsic C1r2-C1s2.     

S phase, an evolutionary chromatin condensation state from G1 to G2, in a breast epithelial cell line.

In order to better understand the changes in DNA organization during the cell cycle, we quantified the chromatin texture of breast epithelial cells and followed its evolution through a cell cycle. The diversity of quiescent cell states led us to limit this study to proliferating cell phases, and to choose a cell line with no G0 cells, the MDA AG cell line. We recently developed a methodology for characterizing in situ the cell cycle of breast epithelial cell lines using a cell image processor. This method is based on 15 densitometric and texture parameters computed on individual Feulgen-stained nuclei and on multiparametric analysis of the resulting data. Chromatin pattern assessment is based on nine texture parameters measured from grey-level co-occurrence and run-length section matrices. In the present study, texture parameter computation showed gradual and progressive modifications of nuclear texture. While discrimination of G1, G2 and M phases was possible, we could not discriminate G1 from S and S from G2. The chromatin pattern (defined by these nine parameters) in the G1 and early S phases, on the one hand, and in the late S and G2 phases, on the other hand, were similar. The parameter values of cells in the S phase progressively increased from G1 to G2. Two interphase chromatin condensation states were distinguished in these breast cells: a base state characteristic of a prereplicative stage and a very granular state characteristic of a postreplicative stage. We hypothesized that S cells are a blend of these two states, the evolution of a non-duplicated state toward a duplicated one.