Enhancing cardiovascular artificial intelligence (AI) research in the Netherlands: CVON-AI consortium.

BACKGROUND: Machine learning (ML) allows the exploration and progressive improvement of very complex high-dimensional data patterns that can be utilised to optimise specific classification and prediction tasks, outperforming traditional statistical approaches. An enormous acceleration of ready-to-use tools and artificial intelligence (AI) applications, shaped by the emergence, refinement, and application of powerful ML algorithms in several areas of knowledge, is ongoing. Although such progress has begun to permeate the medical sciences and clinical medicine, implementation in cardiovascular medicine and research is still in its infancy. OBJECTIVES: To lay out the theoretical framework, purpose, and structure of a novel AI consortium. METHODS: We have established a new Dutch research consortium, the CVON-AI, supported by the Netherlands Heart Foundation, to catalyse and facilitate the development and utilisation of AI solutions for existing and emerging cardiovascular research initiatives and to raise AI awareness in the cardiovascular research community. CVON-AI will connect to previously established CVON consortia and apply a cloud-based AI platform to supplement their planned traditional data-analysis approach. RESULTS: A pilot experiment on the CVON-AI cloud was conducted using cardiac magnetic resonance data. It demonstrated the feasibility of the platform and documented excellent correlation between AI-generated ventricular function estimates as compared to expert manual annotations. The resulting AI solution was then integrated in a web application. CONCLUSION: CVON-AI is a new consortium meant to facilitate the implementation and raise awareness of AI in cardiovascular research in the Netherlands. CVON-AI will create an accessible cloud-based platform for cardiovascular researchers, demonstrate the clinical applicability of AI, optimise the analytical methodology of other ongoing CVON consortia, and promote AI awareness through education and training.

A non-invasive cardiac output measurement as an alternative to the test bolus technique during CT angiography.

AIM: To investigate the association between a non-invasive cardiac output (CO) measurement and the scan delay, as derived from a test bolus injection protocol. The secondary objective was to determine which factors affect the relationship between the CO and scan delay. MATERIALS AND METHODS: Fifty-five patients referred for a contrast-enhanced (thorax-)abdomen CT examination were included in this feasibility study. A test bolus examination was performed prior to the abdominal CT. During the test bolus injection, the CO of the patient was measured using a non-invasive finger-cuff measurement. Associations were analysed using linear regression analyses. Age, gender, height, weight, and blood pressure were included as potential confounders. RESULTS: Linear regression analysis showed a negative and significant association between CO and delay. The regression formula was as follows: scan delay (seconds) = 26.8-1.6 CO (l/min), with a 95% CI between -2.3 and -1.0 (p<0.001). Weight appeared to be a confounder in this relation, and gender and blood pressure were effect modifiers. There was no interaction between scan delay and age, height and weight. CONCLUSIONS: There is a negative and significant association between the non-invasive CO measurement and the CT scan delay; however, to validate these findings a larger cohort study is needed to investigate whether the non-invasively determined scan delay is as accurate as the use of a test bolus.

Feasibility of a low concentration test bolus in CT angiography.

AIM: To investigate the feasibility of using a low-concentration test bolus in abdominal aorta computed tomography (CT) angiography (CTA). MATERIALS AND METHODS: In 10 patients referred for CTA of the abdominal aorta with a body mass index (BMI) ≤28 kg/m2, a standard test bolus of 10 ml contrast medium (CM; 350 mg iodine/ml) was compared with a low-concentration test bolus (5 ml CM; 350 mg iodine/ml; 1:1 diluted with saline) in terms of time to peak enhancement (tPE) and peak enhancement (PE). RESULTS: No significant differences were found between the standard and low-concentration test bolus in terms of tPE and PE. CONCLUSIONS: A low-concentration test bolus (5 ml, 1:1 diluted with saline) is feasible in patients with a BMI ≤28 kg/m2.

Reducing contrast medium volume and tube voltage in CT angiography of the pulmonary artery.

AIM: To evaluate image quality after contrast medium (CM) and tube voltage reduction in computed tomography angiography (CTA) of the pulmonary artery. MATERIALS AND METHODS: Thirty-three patients referred for CTA of the pulmonary artery for suspected pulmonary embolism were included. Patients were randomly assigned to Protocol I (100 ml of 350 mg iodine/ml iodinated CM; n=16) or Protocol II (50 ml of 350 mg iodine/ml iodinated CM; n=17). Dual-energy CT (80 kV and 140 kV) was performed in all patients. An averaged weighted series equivalent to a 120 kV image acquisition was reconstructed. The mean attenuation value of CM was measured at eight positions in the pulmonary trunk and pulmonary arteries. Signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were calculated. Qualitative assessment of the vascular enhancement was performed independently by two experienced radiologists using a three-point scale. Mean attenuation values, image noise, CNR, and SNR of images with 50 ml CM and images with 100 ml CM were compared and mean attenuation values, image noise, CNR, and SNR in 80 kV images and 120 kV images were compared. For qualitative analysis, interobserver variability was analysed using Cohen's kappa statistics. RESULTS: The mean attenuation values in Protocol I and Protocol II were not significantly different at 80 kV (634.6±168.3 versus 537.9±146.7 HU; p=0.088) and 120 kV (482.8±127.7 versus 410.4±106.0 HU; p=0.085). The mean attenuation value at 80 kV was significantly higher than the mean attenuation value at 120 kV in Protocols I and II (p<0.001). The CNR and SNR were higher at 120 kV than at 80 kV in both protocols (p=0.000-0.019); however, there were no significant differences in the CNR and SNR between both protocols (p=0.600-0.952). Qualitative (subjective) analysis showed no statistical significant difference between Protocols I and II (p=0.524-1.000). CONCLUSION: Low tube voltage (80 kV) CTA using 50 ml CM is not inferior to CTA at 120 kV using 100 ml CM.

Low-dose CT angiography of the abdominal aorta and reduced contrast medium volume: Assessment of image quality and radiation dose.

AIM: To determine the effect of using 80 kV tube voltage and a reduced amount of contrast medium on the image quality and radiation dose of computed tomography angiography (CTA) of the abdominal aorta. MATERIALS AND METHODS: Patients who were referred for a CTA examination of the abdominal aorta were included in this technical efficacy study. Thirty patients were divided randomly into two groups. Fifteen patients underwent a dual-energy CT (DECT) protocol (Group A). Fifteen patients were scanned with the use of an automated tube potential selection algorithm tool (Group B). In both protocols, a test bolus injection of 10 ml ioversol (350 mg iodine/ml) was used, followed by 20 ml of 1:1 saline-diluted contrast medium. Quantitative analysis comprised determination of the mean attenuation and contrast-to-noise ratio. Qualitative image analysis was performed independently by five radiologists. The estimated radiation dose in terms of CT dose index and effective dose was recorded and compared with a standard 120 kV protocol. RESULTS: In Group B, six patients underwent CTA at 80 kV, seven patients underwent CTA at 100 kV and two patients underwent CTA at 120 kV. The mean contrast-enhancement values of Group A (80 kV) and the 80 kV subgroup of Group B were 16.5% and 27.6% higher compared to the 100 kV subgroup of Group B, these differences were, however, not significant. There were no significant differences in mean image quality between groups. In patients undergoing CTA at 80 kV the effective dose decreased by up to 51.3% compared to a conventional 120 kV CTA protocol. CONCLUSIONS: The findings of this study support the hypothesis that 80 kV in CTA of the abdominal aorta can reliably be used with only 30 ml contrast medium in total and a 50% reduction in radiation dose. The overall image quality was diagnostically adequate; however, it appeared to be suboptimal in patients with a BMI above 28 kg/m(2).

Isolation and characterization of the erythroid progenitor cell: CFU-E.

Erythroid progenitor cells, CFU-E (colony-forming-unit-erythroid), were isolated to practical homogeneity by a combination of three enrichment procedures. CFU-E were generated in large amounts in spleens of mice previously bled and treated with the erythropoiesis-suppressing drug thiamphenicol. The average CFU-E concentration in spleens from mice 4 d after the thiamphenicol-treatment was 10%. These CFU-E were separated from lymphocytes, erythrocytes, and granulocytes and their progenitor cells by centrifugal elutriation and Percoll density gradient centrifugation. A three- to five-fold enrichment was obtained by elutriation, leading to a CFU-E concentration of 45%. With the Percoll gradient another twofold enrichment was achieved, providing us with a 80-100% CFU-E cell population. The overall recovery of CFU-E was 60-70%. This is a cheap, rapid, and highly efficient method of obtaining large quantities of viable CFU-E. The sequential formation of two-, four-, and eight-cell colonies from CFU-E cultured in vitro was studied. These cells enable us to study the biochemical changes occurring in the differentiation process of an erythroid progenitor cell induced by the hormone erythropoietin. The morphological and some physical and biological properties of these cells are presented.

Hemotoxicity by prolonged etoposide administration to mice can be prevented by simultaneous growth factor therapy.

In this study, we determined in vivo interactions between hemopoietic growth factors and etoposide (VP-16) to assess whether normal blood cell production could be maintained during chemotherapy if hemopoietic growth factors were simultaneously administered. Groups of mice were treated for 7 consecutive days with four different doses of VP-16 in combination with three different doses of erythropoietin (EPO) or granulocyte colony-stimulating factor (G-CSF). In total, 12 combinations of VP-16 plus EPO and 12 combinations of VP-16 plus G-CSF were thus evaluated. Intricate dose-response surfaces of the effects of the different treatments on colony-forming units-erythroid, reticulocytes, hematocrit, colony-forming units-granulocyte/macrophage, and absolute neutrophil count were obtained, which revealed that: (a) simultaneous EPO administration was able to maintain reticulocyte production and to protect mice from VP-16 induced anemia; (b) simultaneous G-CSF administration was able to maintain granulocyte production and to protect mice from VP-16 induced neutropenia; (c) VP-16 dose escalation was feasible when EPO or G-CSF were simultaneously administered; and (d) no increased myelotoxicity on erythroid or granuloid progenitors was observed when EPO or G-CSF was simultaneously administered with VP-16. These results suggest that in vivo either individual hemopoietic progenitors can become resistant against VP-16-induced cell death by appropriate simultaneous growth factor administration or that the loss of overall cell amplification, induced by VP-16, can be compensated by extra amplification of surviving progenitors. Furthermore, these data indicate that a strict separation in time of cytostatic drug and growth factor treatment is not necessarily the optimal schedule with respect to the reduction of hemotoxicity.

Inhibition of hemopoiesis in vitro by neuroblastoma-derived gangliosides.

Hemopoiesis is disturbed in bone marrow-involving cancers like leukemia and neuroblastoma. Shedding of gangliosides by tumor cells may contribute to this tumor-induced bone marrow suppression. We studied in vitro the inhibitory effects of murine neuroblastoma cells (Neuro-2a and C1300) and their gangliosides on hemopoiesis using normal murine hemopoietic progenitor colony-forming assays. Transwell cultured neuroblastoma cells showed a dose-dependent inhibition on hemopoiesis, indicating that a soluble factor was responsible for this effect. Furthermore, the supernatant of Neuro-2a cultured cells inhibited hemopoietic proliferation and differentiation. To determine whether the inhibitory effect was indeed due to shed gangliosides and not, for instance, caused by cytokines, the effect of DL-threo-1 -phenyl-2-decanoylamino-3-morpholino-1-propanol (DL-PDMP) on Neuro-2a cells was studied. DL-PDMP is a potent inhibitor of glucosylceramide synthase, resulting in inhibition of the synthesis and shedding of gangliosides. The initially observed inhibitory effect of supernatant of Neuro-2a cells was abrogated by culturing these cells for 3 days in the presence of 10 microM DL-PDMP. Moreover, gangliosides isolated from Neuro-2a cell membranes inhibited hemopoietic growth. To determine whether the described phenomena in vitro are a reflection of bone marrow suppression occurring in vivo, gangliosides isolated from plasma of neuroblastoma patients were tested for their effects on human hemopoietic progenitor colony-forming assays. These human neuroblastoma-derived gangliosides inhibited normal erythropoiesis (colony-forming unit-erythroid/burst-forming unit-erythroid) and myelopoiesis (colony-forming unit-granulocyte/macrophage) to a higher extent compared with gangliosides isolated from control plasma. Altogether these results suggest that gangliosides shed by neuroblastoma cells inhibit hemopoiesis and may contribute to the observed bone marrow depression in neuroblastoma patients.

The transbilayer distribution of phosphatidylethanolamine in erythroid plasma membranes during erythropoiesis.

Fluorescamine was used to assess the transbilayer distribution of phosphatidylethanolamine in the plasma membrane of murine erythroid progenitor cells, CFU-E (colony-forming unit erythroid), at different stages of their differentiation pathway. Intact cells were exposed to increasing concentrations of fluorescamine and the amount of labeled phosphatidylethanolamine was determined by measuring the fluorescence intensity of its fluorescamine derivative. A semilogarithmic plot of the dose-response curve revealed three different pools of phosphatidylethanolamine, representing its fractions in, respectively, the inner- and outer monolayers of the plasma membrane and subcellular membrane systems. These results show that 9-11% of the total cellular phosphatidylethanolamine is present in the outer leaflet and 9-10% of it is located in the inner leaflet of the plasma membrane in early as well as late erythroblasts. This symmetric distribution of phosphatidylethanolamine over the two halves of the bilayer in the plasma membrane of CFU-E is very similar to that observed earlier in the plasma membrane of friend erythroleukaemic cells (Rawyler, Van der Schaft, Roelofsen and Op den Kamp (1985) Biochemistry 24, 1777-1783). These observations imply that the characteristic asymmetric distribution of phosphatidylethanolamine, as is found in mature erythrocytes, is accomplished at a very late stage of erythropoiesis and possibly during enucleation of the cells or shortly thereafter.

1-phenyl-2-decanoylamino-3-morpholino-1-propanol chemosensitizes neuroblastoma cells for taxol and vincristine.

In this study, we show that an inhibitor of glycosphin-golipid biosynthesis, D,L-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol (PDMP), increases the chemosensitivity of neuroblastoma tumor cells for Taxol and vincristine. At noneffective low doses of Taxol or vincristine, the addition of a noneffective dose of PDMP resulted in 70% cytotoxicity, indicating synergy. Such an effect was not observed for etoposide (VP16). PDMP caused an early (6 h) increase in ceramide (Cer) levels, but the excess Cer was metabolically removed in the long-term (96 h). However, upon incubation with PDMP in combination with Taxol, but not with etoposide, Cer levels remained elevated at 96 h. These results suggest that neuroblastoma cells are normally able to metabolically remove excess Cer, but lose this capacity upon exposure to microtubule modulating anticancer agents (Taxol or vincristine). In addition, PDMP treatment resulted in a decreased efflux of [14C]Taxol and [3H]vincristine from neuroblastoma cells, similar to treatment with PSC833 or MK571, suggesting an effect of PDMP on the transporter proteins P-glycoprotein and/or multidrug resistance protein. PDMP did not further reduce [14C]Taxol or [3H]vincristine efflux in PSC833-treated cells, although it did further diminish cell survival under these conditions. We conclude that a combined administration of nontoxic concentrations of PDMP and either Taxol or vincristine results in highly sensitized neuroblastoma cells. This appears to involve a sustained elevation of Cer levels, possibly in concert with increased drug accumulation.

A new system for the study of erythroid cell differentiation.

Highly purified erythroid progenitor cells (CFU-E), were cultured in vitro for 48 h. Cell division was monitored regularly and cells were isolated at times when maximal colony formation occurred. The cell mixture after one and two cell divisions consisted mainly of proerythroblasts and basophilic erythroblasts (early erythroid cells). After four cell divisions, predominantly polychromatic and orthochromatic cells (late erythroid) were present. After two days of culture, the majority of inoculated cells had differentiated into reticulocytes. This time course in vitro was comparable to the differentiation of CFU-E in vivo. As judged by cytochrome-c oxidase activity, there was an increase in total mitochondrial activity per culture; however, the activity per cell decreased during the differentiation process. The formation of hemoglobin started after the first cell division. This system is presented as a new tool for the study of biochemical processes accompanying erythroid cell differentiation.

The regeneration of stem cells after a bone marrow depression induced by thiamphenicol.

Thiamphenicol (TAP) administered subcutaneously for 4 days via a dialysis bag to bled mice, caused a total inhibition of the differentiation of BFUE into CFUE. The femoral BFUE numbers remained constant whereas the CFUE compartment was virtually absent. The proliferation of CFUs and BFUE was inhibited as well. The granuloid cell line was hardly affected. Shortly after removal of the drug, proliferation of CFUs and BFUE was resumed and a specific time sequence in the regeneration of erythropoiesis in marrow and spleen was observed. A decline in the number in BFUE in the femoral marrow was followed by an increase in the total CFUE in the marrow and an increase in the number of BFUE in the femoral marrow was followed by an increase in the total CFUE in the marrow and an increase in the number of BFUE in the spleen. Maximal erythropoiesis in the spleen occurred 1-2 days later than in marrow. The first wave of erythropoiesis in marrow was followed by a second one leading to high normal values of femoral stem cells. Spleen erythropoiesis consisted of one wave of proliferation and differentiation. The experimental system offers a valuable model for studying certain aspects of erythropoiesis and provides good starting material for CFUE isolation.

Long-term recombinant human granulocyte colony-stimulating factor (rhG-CSF) treatment severely depresses murine marrow erythropoiesis without causing an anemia.

We hereby report profound effects of long-term granulocyte colony-stimulating factor (G-CSF) administration on murine erythropoiesis. Recombinant human (rh)G-CSF (150 micrograms/kg body weight/day) was administered over 24 days to female C57Bl mice. Marrow erythroid colony-forming units (CFU-E) and erythroblast numbers declined to less than 5% of normal, whereas splenic erythropoiesis simultaneously increased. Splenic erythropoiesis effectively compensated for the loss of marrow erythropoiesis as indicated by the maintenance of a normal hematocrit. In the marrow the numbers of spleen colony-forming units (CFU-S) and erythroid burst-forming units (BFU-E) declined as well. Simultaneously, however, these numbers increased both in the spleen and in the peripheral blood by a factor of 20 to 30. These findings suggest a continuous migration of stem cells and progenitor cells out of the marrow and an efficient seeding in the spleen, directly or indirectly induced by G-CSF. In addition the differentiation and/or amplification of BFU-E to CFU-E was impaired in the marrow but not in the spleen. The marrow and splenic microenvironment also behaved differently with respect to granulopoiesis. G-CSF did not lead to an enhanced granuloid amplification in the spleen but exerted its proliferation activity mainly in the marrow. These findings imply that prolonged G-CSF treatment might cause erythroid depression in animals and humans when spleen erythropoiesis is less efficient than in mice.

Induction of globin mRNA transcription by erythropoietin in differentiating erythroid precursor cells.

The effects of erythropoietin (epo) on the proliferation of late erythroid progenitor cells (CFU-E) and on the formation of hemoglobin and of globin mRNA in these cells are described. CFU-E were isolated from thiamphenicol-pretreated anemic mice by elutriation and Percoll density gradient methods. These CFU-E are restricted in their capacity to proliferate in vitro without added epo. The epo dependence in vitro was not absolute. With no epo in the culture medium the first cell division was unimpaired, whereas the third division was only 1%-2% of the control. In the absence of epo the synthesis of hemoglobin is very low in CFU-E, but is increased significantly after about 5 h of incubation with epo present. In epo deprived cells there was considerable hemoglobin formed at about 14 h, but not earlier. The presence as detected by the Northern blot technique of globin mRNA, isolated from CFU-E, was variable, probably depending on the presence of some more mature erythroid cells. By an extrapolation method we show evidence that pure CFU-E would have virtually no detectable globin mRNA. The production of globin mRNA is rapidly (2 h) induced in cells incubated with epo. We conclude that epo, besides having a mitogenic effect on CFU-E, induces the rapid expression of the globin genes.

Expression of the 5'-flanking region of the beta major-globin gene in cultured murine erythroid precursor cells.

The amounts of mouse beta major-globin mRNA and globin gene upstream RNA polymerase III transcripts were compared during the differentiation of purified erythroid colony-forming progenitor cells (CFU-E) in vitro. The accumulation of each RNA was determined relative to the 100% levels in fetal liver erythroid cells. The mRNA level was low in freshly isolated CFU-E and began to accumulate only after a lag period of approximately 2-4 h. It continued to accumulate for approximately 40 h thereafter, at which point it was comparable to that in the fetal liver. In contrast, in three of four CFU-E preparations, the relative level of upstream RNAs was high in freshly isolated CFU-E and reached the maximal level (defined as 100% of the fetal liver level) more rapidly than did the mRNA. The early induction and accelerated accumulation of upstream RNAs in immature erythroid cells suggest some role for these RNAs at an early stage of globin gene activation.

Migration of stem cells and progenitors between marrow and spleen following thiamphenicol treatment of mice.

Recovery of hemopoiesis was studied after a 3-day treatment with the antibiotic thiamphenicol (TAP). A contrasting behavior of the spleen colony-forming units (CFU-S), granulocyte-macrophage colony-forming units (CFU-GM), erythroid burst-forming units (BFU-E), and erythroid colony-forming units (CFU-E) numbers in the bone marrow versus those in the spleen was found. Whereas the cell numbers reached nadirs in the marrow, they peaked 30 to 100-fold above control values in the spleen on day 4. Simultaneously the number of CFU-S, BFU-E, and CFU-GM, but not of CFU-E, increased drastically in the peripheral blood. The tritiated thymidine kill of the splenic CFU-S was too small to explain the endogenous splenic production of these cells. A quantitative analysis further revealed that an effective erythropoiesis was established in the spleen. As a consequence, the first part of a reticulocytosis was mainly due to the splenic contribution, whereas the second part predominantly originated from a delayed marrow erythropoiesis. In contrast, the CFU-GM of the spleen did not effectively differentiate into granuloid precursors. The bulk of the granuloid production occurred in the marrow. The best explanation for these results is a net migration of CFU-S, BFU-E, and CFU-GM from the marrow to the spleen during early recovery, and a back-migration of CFU-GM to the marrow later in the recovery phase. These observations indicate a link between migration of hemopoietic cells and their differentiation at the two hemopoietic sites.

Hemopoiesis during thiamphenicol treatment. I. Stimulation of stem cells during eradication of intermediate cell stages.

Continuous treatment of C57bl/6 mice for 4 days with the cytostatic antibiotic thiamphenicol revealed a dual response of hemopoietic cells. On one hand, morphologically recognizable erythroid precursors and late progenitors (erythroid colony-forming units; CFU-E) and, to a lesser extent, granuloid precursors were found substantially reduced. On the other hand, early granuloid (granulocyte-macrophage colony-forming units; CFU-GM) and erythroid (erythroid burst-forming units; BFU-E) progenitors increased on day 3 to 220%-240% and 120%-130% of the control value, respectively. This was accompanied by a decline of the initial spleen colony-forming units (CFU-S) (day 8) pool size to approximately 60%. These patterns were similar in the bone marrow and the spleen. In addition, the tritiated thymidine kill of femoral and splenic CFU-S rose significantly (p less than 0.05) from 16% +/- 3% to 38% +/- 2% and from 3% +/- 1% to 17% +/- 2%, respectively. A sudden decline of peripheral reticulocytes between days 2 and 3 from 2.8% +/- 0.3% to 0.6% +/- 0.2% was observed, whereas the hematocrit gradually decreased from day 1 to day 4 from 45.2% +/- 0.1% to 39.3% +/- 0.3%. The white blood cells were not affected. From these results we conclude that stem cells were stimulated as a consequence of the suppression of the intermediate cell stages. As analyzed in the accompanying paper, this confirms a prediction stated by a quantitative theoretical concept of in vivo stem cell regulation.

Hemopoiesis during thiamphenicol treatment. II. A theoretical analysis shows consistency of new data with a previously hypothesized model of stem cell regulation.

In a recent theoretical model of stem cell regulation, specific quantitative assumptions were made about an in vivo feedback process from erythroid and granuloid precursor cell stages to the spleen colony-forming units (CFU-S), erythroid burst-forming units (BFU-E), and granulocyte-macrophage colony-forming units (CFU-GM). Utilizing specific effects of the antibiotic thiamphenicol (TAP), new experiments have been performed to challenge this model. Here these data are treated in an analysis that implies three steps. First, model assumptions about TAP toxicity are justified. The toxic TAP effects on erythroid and granuloid precursors are quantified as a continuous reduction of the normal amplification coefficient for CFU-E (down to 1/250), proerythroblasts, basophilic erythroblasts, and proliferating granuloid precursors (down to 1/4). Second, the original model predictions for the behavior of CFU-S, CFU-GM, and BFU-E are compared with the corresponding data. Third, discrepancies are discussed and it is demonstrated that adjustment of one single parameter resolves most of them. Thus one can quantitatively explain the experimental results for CFU-S, BFU-E, and CFU-GM by an activation of the regulatory process postulated: the decline in erythroid (and granuloid) cell numbers enhances the cycling of CFU-S while their self-renewal probability is reduced; consequently CFU-S numbers decline; as more cells differentiate towards BFU-E and CFU-GM per unit time the cell numbers of these cell stages increase. Thus the new data on stem cell behavior during TAP treatment support the hypothesis of a feedback from erythroid and granuloid precursors to the stem cells.

Mechanistic options of erythropoietin-stimulated erythropoiesis.

The in vivo mechanism of hematopoietic growth factor-induced cell multiplication is in debate. Several options can be examined: 1) growth factors can reduce the cycling time of their dividing target cells, 2) growth factors can add extra cell divisions within the differentiation pathway, 3) the combination of the first two possibilities, and 4) growth factors can prevent premature cell death (apoptosis) from occurring in the absence of the stimulating factor. We studied these options in vitro and in vivo in the murine erythroid pathway. Results from in vitro cultures of purified splenic colony-forming units-erythroid (CFU-E), with and without erythropoietin (Epo), and in vivo Epo treatments of thiamphenicol (TAP)-pretreated mice showed neither reduction in cycle times nor addition of extra cell divisions in the differentiating erythroid lineage. The phenomenon of apoptosis was demonstrated as time- and Epo-dependent in vitro with electrophoretic (DNA-ladder), flow-cytometric (subdiploid cells), and morphologic (fragmented nuclei) methods applied on CFU-E. A high dose of Epo administered to mice caused a rapid transient rise in the number of CFU-E to 350% of normal. Early erythroblasts also increased, whereas burst-forming unit-erythroid (BFU-E) numbers did not change. Our results favor a mechanism in which Epo acts as a survival factor for early erythroid cells (CFU-E and early erythroblasts) in vitro, as well as in vivo, preventing apoptosis.

Optimal erythroid cell production during erythropoietin treatment of mice occurs by exploiting the splenic microenvironment.

In this study, quantitative effects on erythroid cell production by a prolonged recombinant human erythropoietin (rhEpo) treatment of mice are presented. Epo treatments, given subcutaneously (s.c.) twice per day in doses of 0.5 to 500 U per day, were performed under steady-state production conditions. We found striking differences between the behavior of the different erythroid cell compartments (burst-forming unit erythroid [BFU-E], colony-forming unit erythroid [CFU-E] and erythroid precursors), as well as between the microenvironments of bone marrow and spleen. Whereas the total-body BFU-E was not changed by Epo, a redistribution of BFU-E from marrow to spleen occurred, resulting in decreasing marrow and increasing splenic BFU-E numbers. Splenic BFU-E produced CFU-E as much as 8 times more efficiently than marrow BFU-E at 50 U of Epo. At low Epo doses (to 1 U/day) no difference was found. The CFU-E in the spleen produced erythroblasts at a higher efficiency at all Epo doses (1.5 to 5 times). It seems as if this efficiency was higher at low Epo doses. Because of the migration phenomenon and the excellent microenvironment in the spleen, at the highest Epo concentrations nearly 70% of all erythroid cells reside in the spleen. Even at the highest Epo doses, granuloid cell production was not affected. Similar to the BFU-E, total-body granuloid cells remained constant (despite a shift of granulocyte-macrophage progenitors [CFU-GM]) from marrow to spleen; however, these cells did not flourish in the spleen. Under these conditions, 90% of the granuloid precursors were still localized in the marrow. Erythropoietin did not change the transit time of erythroid cells at high Epo doses.