| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print December 23, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Division of Immunogenetics, Department of Pediatrics, University of Pittsburgh School of Medicine, Pennsylvania
1 Correspondence: University of Pittsburgh, Pediatrics, 3460 Fifth Avenue, Pittsburgh, PA 15213. E-mail: luppip{at}pitt.edu
In the paper "Monocytes are progressively activated in the circulation of pregnant women" [1 ], we reported the presence of changes in the expression of adhesion molecules and intracellular cytokine production in monocytes of pregnant women followed longitudinally throughout gestation. The most remarkable changes were observed in late pregnancy and were consistent with the concept first introduced by Sacks et al. in 1998 [2 ] that normal pregnancy is associated with a generalized inflammatory response.
Faas et al. now address the questions of why not all the data about adhesion molecule expression are consistent with ours [3
] and other previously published data [2
]; how we can detect intracellular interleukin (IL)-12 and tumor necrosis factor 
(TNF-) production in unstimulated monocytes of pregnant women; and why intracellular IL-12 production in monocytes of pregnant women decreased after endotoxin or phorbol 12-myristate 13-acetate (PMA)/ionomycin stimulation. In Veenstra van Nieuwenhoven et al. [4
], a significant, spontaneous production of IL-12 and TNF- in monocytes was not confirmed.
In our previously published paper [3 ], we focused on the analysis of leukocyte phenotypic characteristics in women during third-trimester gestation. The results of our study are not only in agreement with those of Sacks et al. [2 ] but are also consistent with the findings of another group [5 ]. Specifically, all three groups found a significant increase in the expression of the CD11b antigen on monocytes of pregnant women as compared with nonpregnant controls. Our group, together with Sacks et al. [2 ], also reported a significant increase in CD11b expression on granulocytes during third-trimester pregnancy. In addition, all reported a significant increase in the expression of the CD64 antigen on monocytes of pregnant women. Sacks et al. [2 ] also showed an increase in the expression of CD64 on granulocytes of pregnant women, a finding that we could not corroborate, as the analysis of this marker was not performed in that leukocyte subset in our own study. We described changes in the expression of CD11a and CD54 antigens on the surface of circulating monocytes during pregnancy. These antigens were not included in the list of leukocyte antigens tested by Sacks et al. [2 ]. The only difference between our study [3 ] and that of Sacks et al. [2 ] was that we reported significant changes in the expression of CD49d and CD62L antigens on monocytes. This difference could be explained in several ways. First of all, the criteria for considering data statistically significant were different in the two studies. Sacks et al. [2 ] considered a statistically significant difference in surface antigen expression for P values <0.01. In our study, data were considered significant for P values <0.05. Therefore, differences in the expression of certain leukocyte antigens (i.e., CD49d) between pregnant and nonpregnant women could be detected as statistically significant in our own study but not in the one from Sacks et al. [2 ]. It is interesting that Naccasha et al. [5 ] also found an increase in the expression of CD49d on monocytes of pregnant women during the third trimester, a finding consistent with ours [3 ].
Another important factor when comparing the results published in our manuscript with those of Sacks et al. [2 ] concerns the cohort of pregnant women analyzed. Although the third-trimester pregnant women tested by Sacks et al. [2 ] were not in labor at the time of the blood draw, our pregnant population included a subset of women who were in active labor at the time of sample collection [3 ]. Indeed, it has been recently demonstrated that further activation of circulating monocytes and granulocytes is associated with the onset of labor [5 ]. This latter explanation can also easily explain the apparent discrepancy between data of adhesion molecule expression in our current [1 ] and previously published manuscript [3 ], as stated by Faas et al. The pregnant women followed longitudinally throughout gestation in our current manuscript [1 ] were, in fact, not in labor at the time of blood collection. We are specifically addressing the effect of active labor on the phenotype of circulating leukocytes in our ongoing study of women followed longitudinally from 28 weeks of gestation to delivery.
The intracellular cytokine data obtained after activation with PMA/ionomycin and/or lipopolysaccharide (LPS) are for a combination of cytokines induced during a short incubation time (4 h in our experimental setting) and of cytokines that have already accumulated in the cytoplasm of the test subject. In our study, we have used an unstimulated control to assess the level of residual cytokine synthesis from in vivo activation. This unstimulated control consisted of whole blood cultured for 4 h with brefeldin A, a potent inhibitor of intracellular Golgi-mediated transport, in the absence of PMA/ionomycin or LPS. We run this unstimulated control for all samples to check for possible individual-to-individual variations of intrinsic production of cytokines. The lower limits of cytokine staining positivity were set by using fluorescently conjugated, irrelevant, isotype-control antibodies, a standard technique also extensively used for analysis of intracellular data [6 , 7 ]. All unstimulated and stimulated cells above this cutoff were considered positive.
As Faas et al. correctly pointed out, the isotype-matched controls are not the same reagent as the antigen-specific antibodies, so that confidence in the similarity of concentration, reagent chemistry, purity, and solubility of protein needs to be assumed and is not always valid, even when using reagents from the same manufacturer. These issues were also our concern, and we therefore make efforts to routinely test each lot of all the antibodies used for flow cytometry (even when purchased from the same company) by doing a titration curve, starting with the concentration suggested by the manufacturer. In accordance with the results of the titration curve, we then use a saturating concentration of each antibody in all our experiments. Isotype-control antibodies are used at a matching concentration to the cytokine-specific antibodies.
In our study of pregnant women followed longitudinally throughout gestation, we found that the limits set by the isotype-matched controls did not always overlap with those of the unstimulated samples for all the cytokines tested. Indeed, we observed a level of intrinsic IL-12 and TNF- in monocytes of women during pregnancy. The percentage of cytokine-positive cells was extremely variable among our cohort of pregnant women. There were women with a percentage of IL-12 and TNF--positive monocytes similar to the one reported by Dr. Faas and co-workers [4
]. Another subset of pregnant women showed a higher percentage of cytokine-positive cells. This may be a significant percentage but not much in the way of level of intracellular cytokine expression, as the mean fluorescence intensity of cytokine staining was always low. We were surprised as was Dr. Faas about these findings. We therefore thought of including an additional control to check the specificity of the cytokine staining. With this in mind, we performed "blocking" experiments in which IL-12 and TNF--positive staining was "blocked" by preincubating the permeabilized cells with purified, unlabeled, anticytokine-specific antibodies, as described in detail by Prussin and Metcalfe [7
]. These unlabeled antibodies were from the same clone, have the same binding properties, and were obtained from the same vendor as the specific, fluorescently conjugated, anticytokine antibody. These experiments showed complete neutralization of the cytokines staining in the samples tested and gave us the same lower limits of cytokine staining as obtained with the use of irrelevant isotype controls (an example is shown in Fig. 8 of the paper). These results, therefore, prompted us to conclude that there is a variable level of intrinsic IL-12 and TNF- in unstimulated monocytes of pregnant women cultured for 4 h with brefeldin A but in the absence of PMA/ionomycin or LPS. These same women were tested while in the nonpregnant state (before the pregnancy or 10–22 months postpartum) and did not show a significant intracellular staining for IL-12 and TNF-.
Our PMA/ionomycin and/or LPS in vitro system to stimulate IL-12 production in monocytes appears to differ from the other published studies referenced by Dr. Faas. Differences include experimental setting, doses, and serotype of LPS used and duration of in vitro stimulation. All these factors may contribute to the discrepancy of data resulting from investigations performed in different laboratories. In fact, experimental conditions to stimulate IL-12 production in monocytes vary among groups of investigators [6
, 8
9
10
], and suppression of IL-12 synthesis by human monocytes after incubation with LPS has also been reported [11
]. As we were able to fully activate monocytes to synthesize other inflammatory cytokines (i.e., IL-1ß, TNF-, IL-6, IL-8) after PMA/ionomycin and/or LPS stimulation, our inability to detect an increase in IL-12-positive monocytes could be explained by the fact that our whole-blood cell-culture conditions were not specifically optimized for inducing maximum IL-12 production in monocytes. Sacks et al. [12
], for example, stimulated Ficoll-isolated peripheral blood mononuclear cells with recombinant interferon-
and LPS at different concentrations for 6 h. The purpose of our investigation was to analyze the ability of leukocytes of pregnant women to synthesize a wide range of inflammatory cytokines, not just IL-12.
Although research in the field of pregnancy immunology is growing, contradictory results are still present in the literature regarding the modifications in number and phenotypes of circulating leukocytes and the T helper cell type 1 (Th1)/Th2 paradigm of cytokine production. Many factors can account for these differences, including different pregnant and control populations, experimental settings, and methodology used. Also, pregestational conditions and genetic factors have to be taken into account, especially for explaining subject-to-subject variations in cytokine production during normal and complicated pregnancy.
Received November 12, 2003; accepted November 13, 2003.
REFERENCES
production J. Exp. Med. 181,537-546Related Articles
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
