Statistics from Altmetric.com
The article of Kunisch et al discussing a cross reactivity of allegedly macrophage-specific anti-CD68 antibodies with fibroblasts and activated endothelial cells demonstrates amply that these antibodies should not be used for the identification of macrophages.1 Yet they have been used for this purpose in nearly all medical disciplines, particularly in vascular diseases. In 1990 we observed that some neointimal cells in experimental transplantation atherosclerosis, human native atherosclerosis, and experimental native atherosclerosis had reacted with both presumptive macrophage-specific antibodies (RAM11, HAM56) and an antibody against muscle actin (HHF35).2 In 1997, Andreeva et al demonstrated that the very same human intimal and neointimal cells were immunopositive, both with anti-macrophage (CD68, HAM56) and anti-muscle actin (asm-1, HHF35) antibodies.3 On the basis of these findings, these authors formed a hypothesis that the macrophage markers involved in these reactions were not indicative of cell histogenesis but of phagocytosis. Neither our observation2 nor the demonstration of Andreeva et al3 had any influence on the practice of macrophage identification by the above mentioned antibodies.
Today, I share Kuhn’s opinion4 that the acceptance or rejection of new scientific ideas depends on their relationship to existing paradigms. If they are in agreement with them they are accepted, but if they contradict them they are usually ignored. When the immunohistochemical identification of macrophages was originally proposed there was no existing paradigm in this field and its authors presented their methods against no substantial opposition. My observation that an unreasonably high amount of macrophages had been identified with new monoclonal antibodies in comparison with previously used electron microscopy was disregarded.5 Rare articles describing the reactivity of the above mentioned anti-macrophage antibodies with other cell phenotypes in other medical disciplines were also neglected.
Kuhn described the scientific process as a conflict, in which less satisfactory paradigms are replaced successively by better ones.4 There is only one way which guarantees the correctness of individual paradigms: a strict observance of the facts. For example, an immunological injury induces an intimal thickening composed only of “macrophages” identified by “macrophage-specific” antibodies in a hypercholesterolaemic rabbit. Serial sections show, however, that the cells in question are smooth muscle cells manifesting both macrophage and muscle actin antigens.2,6 Because macrophages cannot produce muscle actin, the cells must be smooth muscle cells phagocytising lipids, and the paradigm of macrophage-specific antigens should be replaced by the paradigm of phagocytotic antigens.
In the article by Kunisch et al,1 it would be interesting to know whether the extent of the overlap between “macrophage” and fibroblast markers in individual patients correlates with some measures of their rheumatoid arthritis, such as synovium hypertrophy, pannus formation, cartilage erosion, and bone destruction. Also, is there a relationship between anti-CD68 positive synovial fibroblasts and contingent dyslipidaemias in rheumatoid arthritis? In vascular diseases, “macrophage-specific” antibodies react with smooth muscle cell phagocytising lipids.3 A similar process in which phagocytising synovial fibroblasts would become immunopositive with anti-CD68 antibodies may take place in rheumatoid arthritis.
We sincerely thank Dr JT Beranek for his thoughtful letter and his comments on our report. He supports our view that CD68 is not a specific marker for macrophages but rather an antigen indicative of phagocytosis,1 as also expressed in several studies in atherosclerosis and other areas.2–8 Our own continuing experiments also support an interrelationship between phagocytosis and the expression of CD68 proteins. After phagocytosis of conventional phosphatidylcholine/phosphatidylglycerol/cholesterol liposomes (24 hours), the human monocytic cell line THP-1 increased the expression of the CD68 epitope recognised by the monoclonal antibody (mAb) EBM11, but not the CD68 epitope recognised by the mAb KP1 (fig 1). In contrast, only marginal effects were seen in human synovial fibroblasts at this time.
To determine whether this finding is based on redistribution,7 conformational change, or altered glycosylation pattern8 of the CD68 molecule, or on (trans)differentiation of the cells,6 requires further study. The observation that different types of non-macrophage-like cells express the “macrophage” marker CD68 in several diseases, clearly has the consequence that these “macrophage-like” cells have to be more thoroughly identified using other cell-type specific markers and the appropriate technique and fixation. We also agree with Dr Beranek’s point that the revival of morphological or ultrastructural techniques in connection with modern immunohistology/in situ hybridisation is essential in clarifying some of these controversial findings.
As regards the correlations between the extent of overlap between “macrophages” and fibroblast markers in individual patients and their measures of clinical disease or the contingent dyslipidaemias in rheumatoid arthritis, we can only provide a partial answer. When patients with rheumatoid arthritis and osteoarthritis were analysed together by the Spearman rank correlation, the percentage of synovial fibroblasts positive by FACS staining for the anti-CD68 mAbs KP1 or EBM11 showed a significant negative correlation with disease markers such as the number of fulfilled American Rheumatism Association criteria9 or the number of leucocytes in peripheral blood (maximum rs = −0.715; p = 0.006; n = 13). Also, the percentages of synovial fibroblasts positive for the KP1 or EBM11 epitopes of CD68 showed a highly significant positive correlation with each other (rs = 0.951; p = 0.000; n = 13), as well as with other fibroblast markers like Thy-1 (CD90) or prolyl-4-hydroxylase (maximum rs = 0.750; p = 0.002; n = 14).
Finally, we thoroughly agree with Dr Beranek’s thoughtful considerations on the interrelationship between ignoring new scientific evidence and the persistence of incorrect paradigms. His recommendation to return to the strict observance of facts, indeed an incontrovertible basis for scientific conduct, should encourage a discussion on the influence of the human factor10 in scientific peer review.