[Prospective one-year epidemiologic longitudinal study of air pollutants and the incidence of croup]

Among a population of 18,175 children below 7 years of age in medium sized towns and rural areas in south-western Germany 552 (3.03%) cases of croup were registered during a 12 months period in 1984-85 by their physicians. Distributions according to months, sex and age at the onset of the disease were the same as in other recent investigations: The level of the air-pollution measured (SO2, NOx, CO, CO2, ozone and dust) was low (highest monthly means in microgram/m3: SO2 88, NO2 73, NO 119, dust 41). There was no relevant influence of the degree of air pollution on croup-frequency. The rise of croup-frequency shortly after a period of several days of higher pollution was accompanied by an influenza epidemic as proved by virus isolations.

[Measurement of respiratory resistance in healthy persons in the winter of 1985/86 and in January/February 1987]

From November to May 1986 oscillatory resistance Ros was measured 6 times in 72 healthy persons in Cologne. Every person was examined at a fixed time of day, respiratory symptoms and bronchopulmonary diseases were considered. Seasonal fluctuations of resistance were found, with slightly elevated values in autumn and spring compared to the winter. Regression analysis showed a decrease of the mean and range of resistance with increasing height and a slight increase in smokers. There was no relation between the seasonal pattern of resistance and the air pollutants, as expected due to relatively low ambient concentrations. When in January 1987 ambient pollution increased, the investigation was repeated on Jan 21, 1987, the day with the highest concentration (daily average of SO2: 256 micrograms/m3). In February control investigations were performed. The resistance was significantly elevated in January 1987 compared to February 87 or to January 86. These results agree with those of other authors, who found a slight impairment of lung function in healthy persons when daily averages of SO2 concentrations exceeded 250 micrograms/m3. However, changes of the same or a greater extent may be caused by meteorological factors.

Plasma exchange and immunosuppression in rapidly progressive glomerulonephritis: a controlled, multi-center study

In a randomized study of 26 patients with histologically confirmed rapidly progressive crescentic glomerulonephritis, 12 patients were treated with immunosuppressants alone (corticosteroids, cyclophosphamide and azathioprine) while the other 14 patients received not only the identical immunosuppressive treatment but also plasma exchange therapy for four weeks. No statistically significant difference was found between the two groups. After 8 weeks, 73% and 69% of the patients in each respective group showed recompensation of renal function; serum creatinine fell from initially 7.0 and 6.2 mg/dl mean to 2.7 and 2.3 mg/dl mean, and under continued immunosuppression did not rise in the following months. Thus, in non-autoantibody induced rapidly progressive glomerulonephritis, kidney function could be improved substantially by immunosuppressive therapy, but an advantage of supplementary plasma exchange could not be shown.

Intestinal cell proliferation. I. A comprehensive model of steady-state proliferation in the crypt

Cell replacement in the crypt of the murine small intestine has been studied and modelled mathematically under steady-state conditions. A great deal of information is available for this system, e.g. cell cycle times, S phase durations, the rate of daily cell production, the Paneth cell distribution etc. The purpose of the present work was to consider simultaneously as much of these data as possible and to formulate a model based upon the behaviour of individual cells which adequately accounted for them. A simple mathematical representation of the crypt has been developed. This consists of sixteen stem cells per crypt (TC = 16 hr, TS = 9 hr), and four subsequent transit cell divisions (TC = 11 to 12 hr, TS = 8 hr) before maturation. Experimental data considered to test the modelling were LI and data on the number of vertical runs of similarly labelled cells. All data were obtained from the ileum after 25 microCi [3H]TdR given at 09:00 hours. A number of alternative assumptions have been considered and either accepted or rejected. Two alternative model concepts of cell displacement explain the data equally well. One is dependent upon strong local cell generation age determinance while the other could accommodate any weak local cell displacement process in conjunction with an environmental cut-off determinant at the middle of the crypt. Both models provide new interpretations of the data, e.g. certain rates of lateral cell exchange between neighbouring columns (250 to 350 per crypt per day out of a total of 420 cell divisions per day) can be concluded from run data, while LI data provide information about the mechanisms involved in maintaining a position-related age order in the crypt.

Analysis of the changes in the proportion of clustered labelled cells in epidermis

A new cell kinetic approach is presented from which the duration of the S and G2 + M phases can be estimated. The technique involves an analysis of the spatial distribution of labelled cells in sections or sheets of epithelium (i.e. an analysis of clustered labelled cells). The technique is largely independent of the absolute number of labelled cells and hence is not influenced by factors which affect the absolute number of labelled cells. The technique is described and experimental data from dorsal murine skin are presented. The technique has also been simulated mathematically so that the phase durations and their variances could be estimated. The advantages of the technique are: it is technically simple; it provides at least two independent estimates of the phase durations; unlabelled cells need not be counted (compare with LI or PLM analysis); it is independent of variations in the absolute yield of labelled cells, and it is applicable if the LI is low and the S phase is short (where the PLM technique tends to fail).

Growth dependent cell proliferation kinetics of a human malignant melanoma grown in nude mice

A human malignant melanoma was grown in nude mice. The tumour showed an exponential growth for several weeks which gradually slowed down, until following week 8 the tumour growth ceased. The reasons for this growth pattern were examined by labelling techniques (PLM, LI). The tumour cell production as quantified by the growth fraction, showed only a moderate reduction which, taken alone, could not explain the growth cessation. The important mechanism seems to be an increased loss of tumour cells during the intermitotic interval. While the loss of cells/h remains constant the total intermitotic cell loss is increased because the cell cycle times are prolonged by 50% from the exponential phase (45 h in week 3) to the plateau phase (66 h in week 8).

Cell kinetic studies in the epidermis of mouse. III. The percent labelled mitosis (PLM) technique

Full PLM curves have been obtained for four sites in the mouse. The first peaks have been analysed by computer and the duration of the G2 + M and S phases determined together with their standard deviations. The full curves showed a general similarity for all four sites with no clear second peak. The data are compared with the published data for mouse and human epidermis using the in vivo PLM technique. The timing and shape of the first peak can vary considerably even for one site in mice. Hence, both G2 + M and S can vary in their durations. Cells labelled at one time of day exhibit different kinetic properties to those labelled at another time of day. The duration of G2 + M is shortest in dorsum labelled at 03.00 hours (3 X 2 hr) and longest in tail (up to 7 X 5 hr). The S-phase is shortest in dorsum (6 X 3-7 X 2 hr) and longest in tail or ear (13 X 3-14 X 1 hr). There is also a very large standard deviation in tail and foot. There is little general variability when the psoriatic human data are considered, which is surprising. The general variability amongst the data from experimental mice might also be expected amongst humans which might make comparisons between the cell kinetics of normal and diseased skin difficult.

The kinetics of hematopoietic stem cells during and after hypoxia. A model analysis

A previously described mathematical model of the hematopoietic stem cell system has been extended to permit a detailed understanding of the data during and after hypoxia. The model includes stem cells, erythroid and granuloid progenitors and precursors. Concerning the intramedullary feedback mechanisms two basic assumptions are made: 1) The fraction "a" of CFU-S in active cell cycle is regulated. Reduced cell densities of CFU-S, progenitors or precursors lead to an accelerated stem cell cycling. Enlarged cell densities suppress cycling. 2) The self renewal probability "p" of CFU-S is also regulated. The normal steady state is described by p = 0.5, indicating that on statistical average each dividing mother stem cell is replaced by one daughter stem cell, while the second differentiates. Diminished cell densities of CFU-S or enlarged densities of progenitors and precursors induce a more intensive self renewal (p greater than 0.5), such that the stem cell number increases. The self renewal probability declines (p less than 0.5) if too many CFU-S or too few progenitors and precursors are present. The model reproduces bone marrow data for CFU-S, BFU-E, CFU-C, CFU-E, 59 Fe-uptake and nucleated cells in hypoxia and posthypoxia. Although the ratio of differentiation into the erythroid and granuloid cell lines is kept constant in the model, a changing ratio of CFU-E and CFU-C results. The model suggests that stem cells and progenitor cells are regulated by a regulatory interference of erythropoiesis and granulopoiesis.

The kinetics of granulopoiesis in long-term mouse bone marrow culture. Part II

A mathematical model of mouse granulopoiesis in long-term bone marrow culture was constructed, based on established in vivo cell kinetic parameters. We applied the model to the cell kinetic experiment presented in Part I. Comparing model-predicted cell kinetics with the experimental data led to iterative testing of several hypotheses. In the final model, the cell kinetics of intact tissue culture flasks were reconstructed, using the experimental data from 10 days of tube culture. Among other things, our analysis suggests that the parameters of normal in vivo granulopoiesis apply to bone marrow culture.

The kinetics of granulopoiesis in long-term mouse bone marrow culture. Part I

The spontaneous stratification in long-term bone marrow cultures was illustrated and quantified. The cultures were separated into three hematopoietic layers: nonadherent cells in the supernatant medium, lightly adherent cells on top of the stromal layer, and remaining cells buried within the stromal layer. The cells of each layer were subcultured for 10 days in plastic tubes that inhibit the formation of a stromal layer. Daily samplings with absolute and differential cell counts were obtained. We identified three families of cell disappearance curves and cell types: CFU-s, hemocytoblasts, myeloblasts, and promyelocytes (G1, 2); myelocytes (G3); and postmitotic granulocytes (G4). Also, the numbers of mitotic and necrotic cells were determined. The longest half-time of CFU-s was 2.5 days. Lacking stromal support, CFU-s disappeared faster than other differentiated cells. Generally, these cells maintained their numbers for the first week of subcultures, which was attributable to a temporarily maintained balance of cell death and fresh cell production. After more than 7 days, there was a rapid decline of all differentiated cell types.

[Thrombopoiesis, thrombocyte count and thrombocyte function before and following cell separation]

About 50% of the circulating platelets can be taken from healthy donors by means of the cell separation technique. In order to investigate the effect of cell separation on thrombopoesis bone marrow samples were taken from 12 voluntary donors up to 5 times. The megacaryocytes were analyzed before and after cell separation by morphometric measures and determinations of platelet count and platelet function (adhesion, aggregation, spreading, retraction) were carried out. Removing of large amounts of platelets form the circulation by stimulation of thrombopoesis resulted in an increase of the peripherical platelet count already 48 h after cell separation. Of the platelet functions tested only aggregation induced by collagen was found to be increased after cell separation as compared to that before the procedure.

[Mathematical models in hematology]

The use of mathematical models in haematology is shown by some examples concerning stem cell kinetics, erythropoiesis and thrombopoiesis. At first, model assumptions are formulated which include the biological knowledge and some regulatory hypotheses. Then, the reaction of the model on stimulation and suppression is calculated. Finally, by comparison with experimental or clinical data one can evaluate how far the model assumptions are sufficient to understand the measurements. Thus one can exclude wrong hypotheses and identify the important regulatory influences.

Evidence for discrete cell kinetic subpopulations in mouse epidermis based on mathematical analysis

Continuous (repeated) labelling studies in mouse epidermis indicate that nearly all cells are labelled after about 100 hr. Percentage labelled mitoses studies ([3H]TdR at 15.00 and 03.00 hours) have a first peak that does not reach 100% and has a half-width of about 10 hr. Small second and third peaks can be detected at about 90 and 180 hr. respectively. The changes with time in the number of labelled cells show a difference dependent on the time of day of [3H]TdR administration. Both curves show an early doubling in labelled cells which then decline, forming a peak of labelled cells. A second peak occurs at about 120 hr. This is followed by a progressive decline with no further peak until values of about 1% labelling are obtained at 340 hr. These experiments have been investigated mathematically. A computer programme has been devized that permits all three types of experiments to be analyzed simultaneously. More importantly, it can analyse situations with a heterogeneity in cell cycle parameters in all proliferative subpopulations. Various models for epidermal cell replacement have been considered. The data as a whole can best be explained if the basal layer contains at least two distinct subpopulations of cells and an exponentially decaying post-mitotic population with a half-life of about 30 hr. The proliferative sub-populations must be characterized by near integer differences in the length of cycle, the precursor (stem) compartment having the longer cycle. An inverse relationship is required for the length of S, i.e. the shortest time for the stem cells. A full range of cell kinetic parameters can be calculated and are tabulated for the most appropriate model system which is one involving three transit proliferating subpopulations.

A solution to the controversy on stem cell regulation

Supplementary to the discussion in Blood Cells 7 (1981) pp. 417-443 concerning regulatory mechanisms of stem cell proliferation and differentiation, the consequences of the two proposed hypotheses of Blackett and Botnick and Lord have been calculated. It can be shown that their equations are equivalent and can be transformed into each other. However, their verbally described regulatory mechanisms are different and turn out to be more important than the equations. The assumptions of Blackett and Botnick lead to an unstable behaviour of the stem cell system with cyclic cell numbers while the hypothesis of Lord leads to stable recovery curves.

A comprehensive mathematical model of stem cell proliferation which reproduces most of the published experimental results

On the basis of experimental knowledge about haemopoietic stem cells a catalogue of fundamental statements is formulated. From this a simple mathematical model of haemopoietic regulation mechanisms is developed. The functional net effects of regulatory processes which are still unknown or unmeasurable are estimated using evolution arguments. The model is developed in three steps. It allows description of the self-replication of stem cells after direct destruction as well as their reaction to increased or reduced needs in the erythropoietic system. The most important experimental data about changes in CFUs, BFUe and CFUe after acute or chronic irradiation, anaemia, hypoxia, hypertransfusion or direct erythropoietic stimulation can be reproduced within the model. The model allows us to understand most of the results of experimental stem cell research. Furthermore, it can be applied for a more precise analysis of the existing data. Predictions about the results of certain experiments can be made.

A mathematical model of thrombopoiesis in rats

A mathematical model of thrombopoiesis in rats is presented. This has four compartments; stem cells, megakaryocytes, thrombocytes and thrombopoietin. A high thrombopoietin concentration influences bone marrow proliferation in three ways. Firstly the stem cells are stimulated and a slow increase in megakaryocyte number follows. Secondly there are additional endomitoses in the (early) megakaryocytes resulting in an increase in megakaryocyte volume. Thirdly the megakaryocyte maturation time is shortened. The parameters of the model are determined from experimental values for the normal, maximum and minimum proliferation rates, maturation times and destruction rates. The model is tested by comparing simulated results for acute and chronic thrombocytopenia and thrombocytosis with experimental curves from the literature. The model and data agree within the limits of experimental error. Not all of the thrombopoietic regulatory system is known yet, so some important alternative hypotheses are investigated and compared with the model. Several hypotheses have been excluded in this way.

[Computer-assisted diagnosis of haematological disease in hospital patients (author's transl)]

A system of computer-assisted diagnosis is described which, in analogy to conventional diagnosis, takes into account both logical and statistical factors. The arrangement of such a system makes it clear that, after a period of successful use in specialist branches of medicine, a phase of less spectacular success has occurred which, however, provides the conditions for including larger areas of medicine in computer-assisted diagnosis. These systems are likely to conform more closely to the needs of the doctor than previous ones. Whether they can be used in non-hospital practice would depend largely on whether the questions put to the system by the doctor do not take more time and answers can be given in short, easily understandable form. The best type of answer as well as the transmission of information can be discovered empirically. The differential diagnostic level of computer-assisted diagnosis is, according to present experience, quite satisfactory but can be further improved. This development is greatly helped by continuing increases in the speed of calculators and their data storage capacity.