Testing of serum atherogenicity in cell cultures: questionable data published

Abstract

In a large series of studies was reported that culturing of smooth muscle cells with serum from atherosclerosis patients caused intracellular lipid accumulation, while serum from healthy controls had no such effect. Cultures were used for evaluation of antiatherogenic drugs. Numerous substances were reported to lower serum atherogenicity: statins, trapidil, calcium antagonists, garlic derivatives etc. On the contrary, beta-blockers, phenothiazines and oral hypoglycemics were reported to be pro-atherogenic. Known antiatherogenic agents can influence lipid metabolism and cholesterol synthesis, intestinal absorption or endothelium-related mechanisms. All these targets are absent in cell monocultures. Inflammatory factors, addressed by some antiatherogenic drugs, are also not reproduced. In vivo, relationship between cholesterol uptake by cells and atherogenesis must be inverse rather than direct: in familial hypercholesterolemia, inefficient clearance of LDL-cholesterol by cells predisposes to atherosclerosis. Accordingly, if a pharmacological agent reduces cholesterol uptake by cells in vitro, it should be expected to elevate cholesterol in vivo. Validity of clinical recommendations, based on serum atherogenicity testing in cell monocultures, is therefore questionable. These considerations pertain also to the drugs developed on the basis of the cell culture experiments.

Letter to the Editor

Strategies to treat atherosclerosis pharmacologically should take its multifactorial etiology into account [1]. Atherogenesis involves many cell types interacting with each other and with extracellular matrix [2]. Therefore, results obtained in studies on a single cell type should be considered with caution when extrapolated to the whole body. A large series of studies, having become internationally known in 1986 after the publication in The Lancet [3], has been continued until today [4]. Cultures of smooth muscle cells from human aorta and, in some studies, of peritoneal macrophages, were used as an in vitro model for testing of serum atherogenicity and anti- or pro-atherogenic action of various substances. The following, among other things, was reported: within 24 hours of cultivation with diluted (40%) sera of coronary atherosclerosis patients, the total intracellular cholesterol (Ch) increased twofold to fivefold [5]. Low density lipoproteins (LDL) from patients with coronary atherosclerosis caused a twofold to fourfold rise in cholesteryl esters in cultured human blood monocytes and intimal smooth muscle cells isolated from the normal aorta [6], [7]. Cultivation with the sera or LDL from healthy individuals failed to induce intracellular lipid accumulation [5], [6]. Furthermore, calcium antagonists (verapamil, nifedipine, darodipine, isradipine, diltiazem, etc.) reduced the Ch level in cultured cells and, at the same time, lowered the incorporation of ³H-thymidine by the cells, which was interpreted as decreased cell proliferation [8]. It was concluded that calcium antagonists manifested a direct antiatherosclerotic effect in culture [8]. Furthermore, beta-blockers (propranolol, alprenolol, metoprolol, atenolol, pindolol, and timolol) caused a 1.5- to 2-fold rise in Ch level of cultured cells and stimulated their proliferation [8]. Apart from direct admixture of a tested substance to the culture medium, an “ex vivo” model was used: within 2–4 h after an oral administration of a beta-blocker (propranolol), plasma became atherogenic i.e. its addition to the cell culture medium induced intracellular Ch accumulation and stimulated proliferation of the cultured cells. At the same time, blood plasma of patients receiving calcium antagonists acquired antiatherosclerotic properties manifested by its ability to lower the intracellular Ch and to inhibit cells proliferation in a culture [8]. Another example: garlic extract added to the medium, where smooth muscle cells from atherosclerotic plaques had been cultured, significantly reduced the Ch level in the cells and inhibited their proliferation. Blood serum taken from patients 2 hours after an oral administration of 300 mg of garlic powder significantly lowered Ch accumulation in the cultured cells [9]. Remarkably, infiltration of cultured cells by lipids was reported to be associated with their increased proliferation: in the cultures of smooth muscle cells taken from zones of fatty infiltration and fatty streaks in the human aortic intima, the thymidine index exceeded the normal value 4.5- and 3-fold respectively [10]. Pharmacological agents, modifying intracellular lipid accumulation, influenced cell proliferation in the same direction [11]. In general pathology, however, fatty infiltration is seen as a manifestation of cell damage (degeneration) [12], which can hardly be expected to come along with enhanced proliferation.

Furthermore, recommendations for practice were formulated on the basis of cell culture experiments [13], including drug dosage: “To decrease atherogenic potential of serum and to maintain it at a low level, verapamil should be administered at a dose of 40 mg 5 times daily with a 4- to 5-hour interval between doses.” [13] This recommendation should be seen within the scope of a broader question, whether data from cellular models may be used directly for clinical recommendations without clinical trials. Cellular systems in pharmacological research are essential for initial evaluation of drugs. However, their use to predict a response of the whole body is limited. A response of cellular systems to an agent influencing transport across the plasma membrane can have opposite effects in cells compared to the whole body [14]. As an explanation for atherogenicity of serum from patients with coronary atherosclerosis, a causative role of Ch-containing immune complexes was maintained [15]. However, among the mechanisms of atherogenesis induced by immune complexes, discussed in the literature, are inflammatory phenomena such as release of proinflammatory cytokines from macrophages, increase in adhesion molecules and dysfunction of endothelial cells [1], [16], [17], [18], which are not reproduced in the cell monocultures. Known action mechanisms of antiatherogenic or lipid-lowering drugs include regulation of Ch synthesis, lipid metabolism in the liver, intestinal absorption of lipids or endothelial functions [1], [19], [20]. All these targets are absent in the cell cultures. In addition, drugs like statins have numerous “pleiotrophic effects” that can be beneficial [1] but are not reproducible in the cultures. Inflammatory phenomena, addressed by some antiatherogenic agents [2], [21], are also not reproduced by this model. In vivo, the relationship between Ch uptake by cells and atherogenesis is inverse rather than direct. For example, in familial hypercholesterolemia, caused by abnormality of lipoprotein receptors, ineffective clearance of LDL-Ch from serum causes hypercholesterolemia and predisposes to atherosclerosis [22]. Moreover, up-regulation of LDL receptors (and, correspondingly, of the LDL-Ch uptake by cells) is one of the therapeutic strategies for atherosclerosis [23]. Accordingly, if a pharmacological agent reduces Ch uptake by cells in vitro, it should be expected to cause serum Ch elevation in vivo. In other words, pharmacologic agents displaying an “antiatherogenic” effect in cell cultures should be expected to have a pro-atherogenic effect in vivo. Therefore, conclusions and recommendations, formulated on the basis of the cell culture studies discussed above, can be based on misunderstanding. Moreover, an anti-atherosclerotic drug Allicor, a derivative of garlic, was developed on the basis of the cell culture experiments. Anti-atherosclerotic efficacy of Allicor was confirmed by clinical trials [24], [25]. Discussing the mode of action of the Allicor, both articles refer to the cell culture study [8]. In fact, the results of the trials [24], [25] are at variance with those of the cell culture experiments: if garlic indeed lowers the intake of Ch by the cultured smooth muscle cells [9], it might cause elevation of serum Ch in vivo. It is written in [25] with reference to a review [26], obviously implying effectiveness of garlic: “Lipid-lowering properties of garlic-based drugs and preparations are studied rather well” [25]. However, it is stated in [26] that there is increasingly less evidence for lipid lowering properties of garlic preparations. A later review on this topic concluded that evidence, based on rigorous clinical trials of garlic, is not convincing [27]. For hypercholesterolemia, the reported effects of garlic are small and may be of no clinical relevance [27]. Previously we discussed other trials on hypercholesterolemia, results of which have not been convincingly confirmed by other researchers [28]. The matter could be clarified by means of a large-scale independent trial. Note that Allicor is produced by INAT Farma fused with the Institute of Aterosclerosis Research (http://inat.ru), where the above-mentioned cell culture experiments have been performed. With pharmaceutical costs increasing faster than most other health care expenditures, studies should meet the needs of evidence-based treatments and not just the needs of the manufacturers [1].

The material from this paper was presented at several conferences (e.g., [29]); and the author was asked, why it is focused on reports of mostly one research group and the topic is not discussed in a broader perspective. In fact, there have been no other studies, where serum atherogenicity and atherogenic/antiatherogenic potencies of drugs were evaluated directly on cell cultures and practical recommendations given on the basis of such evaluations. Earlier experiments with culturing of smooth muscle cells from atherosclerotic plaques [30], culturing or incubation of different cells in lipoprotein-containing media (e.g., [31], [32]) included neither measurement of serum atherogenicity nor drug evaluations. In conclusion, a reason for submitting this paper to an open access journal is to draw the attention of the scientific community to the fact that doubtful information was published in numerous pharmacological, clinical and other editions (e.g. [7], [10], [12], [33], [34], [35]) including a recently published Russian-language handbook, and is still being referred to in more recent studies (e.g. [36], [37], [38]). However, the study [38] disagreed with the results of some experiments by the research group under discussion in this paper; and an artifact in these experiments was proposed as an explanation for the disagreement.

Notes

Competing interests

The author declares that he has no competing interests.