Intracytoplasmic Detection of Cytokines in Neonatal Lymphocytes and Monocytes by Flow Cytometry

To the Editor:

In a recent publication in BLOOD, Chalmers et al1presented data on cytokine production of lymphocytes at the single-cell level using three-color flow cytometry. A total of 6 blood samples have been investigated. Their main finding was the lower production of interleukin-2 (IL-2), IL-4, interferon-γ (IFN-γ), and tumor necrosis factor-α (TNF-α) by cord blood lymphocytes compared with peripheral blood lymphocytes derived from adults. The investigators concluded that the reduced incidence of graft-versus-host disease in cord blood transplantation may be due to this altered cytokine profile. Performing as well the intracytoplasmic detection of cytokines by flow cytometry on 20 cord blood and 34 adult blood specimens, our results confirm a diminished ability of cord blood lymphocytes to produce IL-2 and IFN-γ (Table 1). However, we found a strikingly lower relative amount of IFN-γ–positive cells (4%v ∼25%1) and a clearly decreased IL-2 expression per cell, suggesting a more pronounced defective cellular immune response in neonates.

These discrepancies can be explained by different technical methods. Chalmers et al1 used isotype-specific antimouse antibodies to discriminate specific versus nonspecific staining. Although these antibodies are essential in the detection of irrelevant specificity of surface molecule staining, they are inadequate negative controls for intracellular staining, because they do not cover the enhancement of unspecific binding during fixation.2 We compared isotype-specific antimouse antibodies to an excess of unlabeled anticytokine antibodies as negative controls (provided by PharMingen, San Diego, CA) in 12 cord blood and 23 adult blood samples. We found that positively and negatively stained cells cannot be exactly differentiated by the use of isotype-specific antimouse antibodies. Because of a high false-positive rate, ranging between 5% and 34% (data not shown), the number of specifically stained cells may be overestimated. We therefore emphasize that only the use of adequate negative controls such as unlabeled anticytokine antibodies enables a meaningful interpretation of data obtained by flow cytometric detection of intracytoplasmic cytokines. The technique of cell separation performing density gradient centrifugation (Ficoll-Hypaque) used in the study of Chalmers et al1 also requires further comment. Although these preparations provide useful information about individual cell types, functional responses of purified cell preparations may not represent physiological reactions occuring in vivo. De Groote et al3 clearly demonstrated that density gradient centrifugation may affect cell activation status, resulting in considerably altered cytokine production compared with whole blood culture technique. Whole blood containing physiological levels of regulatory cytokines may therefore more closely approximate circulating cells in vivo and better reflect disease state alterations. To avoid artificial modifying effects on cell function, we preferred the whole blood assay4 before flow cytometric analysis of intracytoplasmic cytokines.

Additionally, we would like to comment on the decreased production of TNF-α by cord blood lymphocytes.1 This observation could be misinterpreted as a diminished capacity of neonates to produce inflammatory cytokines. However, it should be kept in mind that monocytes are a major cellular source for this cytokine. Andersson et al5 had already shown that cord blood monocytes were able to produce TNF-α at the single-cell level. Moreover, we found an increased amount of IL-6–positive monocytes in cord blood versus adult blood (Table 2). Because IL-6 represents a further step in the lipopolysaccharide-induced cytokine cascade after TNF-α secretion, these results support a well-developed inflammatory reaction in neonates.

In summary, our results based on a larger series show an even more decreased cellular immune response in neonates, characterized by a profoundly reduced production of IL-2 and IFN-γ. However, the ability of neonatal monocytes to produce TNF-α5 and IL-6 (Table2) suggests that an impaired inflammatory response is not responsible for the reduced incidence of graft-versus-host disease in cord blood transplantation. To clarify these and other questions in neonatal immunology, the method presented here has the potential to become the standard assay for examining cytokine production at the single-cell level. Adequate negative controls and the whole blood culture technique instead of cell separation should be preferred to prevent misinterpretation of data.