1-s2.0-S0015028203007787-main

Fraction of the peripheral bloodconcentration of CD56�/CD16�/CD3� cellsin total natural killer cells as an indicationof fertility and infertility

Vassiliki I. Michou, B.Sc.,a,b Panagiotis Kanavaros, M.D., Ph.D.,c

Vassilis Athanassiou, M.D., Ph.D.,d George B. Chronis, M.D., Ph.D.,e Stelios Stabamas,b

and Vassilis Tsilivakos, M.D., Ph.D.b

Locus Medicus Laboratory, Athens, and University of Athens, Zografou, Greece

Objective: To determine whether the peripheral blood concentration of CD56�/CD16�/CD3� cells, the mainnatural killer (NK) subpopulation that colonizes the endometrium in the middle and late secretory phase, canbe related to fertility or infertility status.

Design: A case control study.

Setting: Immunopathology department of an infertility laboratory.

Patient(s): A total 99 women were selected (group I: consecutive spontaneous aborters, n � 25; group II:sporadic spontaneous aborters, n � 30; group III: infertile, n � 33; group IV: controls, n � 11).

Intervention(s): Immunophenotyping of women grouped according to their fertility status.

Main Outcome Measure(s): Peripheral blood lymphocytes were examined by two- and three-color flowcytometry.

Result(s): A statistically significant association between endometrial-type peripheral blood (PB) NK cellconcentrations and fertility status (groups I and IV vs. groups II and III) was documented. The%/TOTAL PB

CD56�CD16�CD3�

NK cells was significantly higher [1] in fertile (groups I and IV) than in sporadicaborters/infertile (groups II and III) women, [2] in group I when compared with groups II and III, and [3] ingroup IV when compared with groups II and III.

Conclusion(s): This study indicates that diminished numbers of the %/TOTAL PBCD56�CD16�CD3�

NK cells are related tosporadic aborters and infertile women. Thus, the fraction could be used as an indicator of subsequentsuccessful implantation and maintenance of gestation. (Fertil Steril� 2003;80(Suppl 2):691–7. ©2003 byAmerican Society for Reproductive Medicine.)

Key Words: Endometrial-type NK cells, abortion, infertility

Natural killer (NK) lymphocytes exert theircytolytic functions through specific cell sur-face molecules that detect target cells that aredeficient in cell-surface class I molecules ofthe major histocompatibility complex (MHC)(1, 2). The peripheral blood NK cells make upabout 4%–10% of the total lymphocyte popu-lation, and on the basis of their antigen expres-sion, three distinct groups can be recognized.Those are CD16�/CD56� and CD16�/CD56�

groups, which represent about 90% of the pe-ripheral blood NK cells, and the CD56�/CD16� group. The latter cells are lesscytotoxic, and studies in biopsies haveshown that in the middle secretory phasethey make up about 70%–90% of the total

lymphocytes found in the endometrium(3, 4).

There is increasing evidence that the ele-vated proportion and activity of all peripheralblood NK cells are related to spontaneous abor-tions, thus functional assays for NK cell cyto-toxicity have been used to monitor treatmentand pregnancy outcome (5–7). IV injectionwith immunoglobulins (IVIg) or allogenic lym-phocyte treatment results in a decreased per-centage and activity of NK cells and in suc-cessful pregnancy outcome, especially inrecurrent aborters (8, 9). These treatments mayowe their success to the production of Th2cytokines (e.g., IL-4, IL-10), which conse-

Received November 19,2001; revised andaccepted March 28, 2003.Supported by LocusMedicus Laboratory,Athens, Greece.Reprint requests: VassilisTsilivakos, M.D., Ph.D.,Locus Medicus Laboratory,53 L. Riankour St., Athens115-23, Greece (FAX:00302106996870; E-mail:[email protected]).a Genetics andBiotechnology Department,Faculty of Biology,University of Athens, Ilissia,Greece.a Department of Infertility.b Department of Anatomy,Histology and EmbryologyMedical School, Universityof Ioannina, Ioannina City,Greece.d Department of IVF,Maternity and SurgicalCentre “Mitera”, Athens,Greece.c Maternity and SurgicalCentre “Mitera”, Athens,Greece.

FERTILITY AND STERILITY�VOL. 80, SUPPL. 2, SEPTEMBER 2003Copyright ©2003 American Society for Reproductive MedicinePublished by Elsevier Inc.Printed on acid-free paper in U.S.A.

0015-0282/03/$30.00doi:10.1016/S0015-0282(03)00778-7

691

quently alter the cytokine Th1/Th2 balance and thus favorpregnancy (10).

Szereday et al. (11) have shown increased IL-2 (Th1)production and reduced IL-10 (Th2) expression in the pe-ripheral blood of recurrent aborters natural killer (NK) cellsexpress IL-2 receptors and during pregnancy proliferate andbecome more potent in the presence of IL-2. Lymphokine-activated killer (LAK) cell activity is more potent than NKcell activity and results from the IL-2 effect on the NK cells.Domination of the Th1 cytokines, which increases LAKactivity, could lead the cytotoxic NK cells, which colonizethe endometrium, to override the presence of the diminishednumbers of MHC molecules C, G, E [human leukocyteantigen (HLA)] expressed in trophoblast cells (12, 13).

A recent study, based on immunophenotypic analysis ofwomen before IVF with a history of infertility has shownelevated CD56�/CD16�/CD3� percentages in the failed group.Elevated NK concentrations in pregnant women with a historyof infertility increased the chances of a miscarriage (14).

In the present study, three-color immunofluorescenceanalysis was performed to depict the immunophenotypiccharacteristics of women, which are characterized by effort-less conception, infertility, or abortions. Emphasis was onCD56�/CD16�/CD3� cells, which in this study are referredto as endometrial-type NK cells since they are immunophe-notypically identical to those found abundantly in the phys-iological endometrium. The concentration of this lympho-cytic population in the peripheral blood was taken intoaccount because it is the best parameter for measuring theiravailability and subsequent supply to the endometrium.

MATERIALS AND METHODS

PatientsFrom November 1999 to March 2001, 780 women experi-

encing fertility problems visited the Immunopathology Depart-ment of the Locus Medicus Laboratory for peripheral bloodlymphocyte immunophenotypic analysis. A full case historyand previous lab tests results were available for assessment. Themajority of the women in the study group suffered from acombination of the following risk factors: tubal obstruction,polycystic ovarian (PCO) disease, endometriosis, recurrent

anovulatory menstrual cycles, autoantibodies, and male factors(Table 1). Given that each of the risk factors contributesdifferently to the individual case history, strict criteria wereused to generate distinct categories according to their fertil-ity history. To meet these criteria, partner-specific women ofthe study group were narrowed down to 88:

Group I.—Consecutive spontaneous aborters [n � 25; meanage, 30.4 � 4.1 years; median (range) of abortions, 3(2–5); mean infertility duration, 2.4 � 1.35 years]:

● With no history of tubal and/or male factor infertility, con-ceiving always within a 3-month period. All conceptions re-sulted in first trimester spontaneous abortions with normalkaryotype.

● With history of tubal and/or male factor infertility, who con-ceived in all (�2) IVF treatments (significant rise in �hCG).All resulted in first trimester spontaneous abortions.

Group II.—Sporadic spontaneous aborters [n � 30; mean age,34.7 � 5.3 years; median (range) of abortions, 2 (1–3);mean infertility duration, 5.03 � 2.81 years]:

● With no history of tubal and/or male factor, one or two firsttrimester spontaneous abortions, after at least a two or threeyear effort, respectively.

● History of tubal and/or male factor, with a less than 25%chance of success in IVF treatments (in four or more at-tempts). All first trimester spontaneous abortions with normalkaryotype.

Group III.—Infertile (n � 33; mean age, 35.4 � 5.4 years; meaninfertility duration, 9.32 � 4.65 years):

● Failed to conceive for the last 5 years or more.● Experienced more than three unsuccessful IVF treatments (no

subsequent rise in �hCG).

Group IV.—Fertile controls (n � 11; mean age, 32.5 � 3.9 years):

● A live birth at least 6–18 months before the study.● Conceived within a 3 month effort in all of their gestations

(one or more).● All gestations resulted in a live birth.

The immunophenotypic tests involving the study group (n �

T A B L E 1

Obstetric information of the study groups (n � 88).

Tubalfactor

Malefactor

Ovulatoryfactor Endometriosis

Male factorand tubal

factor Autoantibodies IVFUnexplained

infertility

Group I (n � 25) 0/25 1/25 0/25 0/25 1/25 0/25 2/25 23/25Group II (n � 30) 2/30 11/30 7/30 9/30 1/30 4/30 12/30 7/30Group III (n � 33) 4/33 5/33 6/33 8/33 2/33 3/33 22/33 5/33

Michou. Endometrial-type NK cells and infertility. Fertil Steril 2003.

692 Michou et al. Endometrial-type NK cells and infertility Vol. 80, Suppl 2, September 2003

88) were carried out 2–5 months subsequent to failed IVF or aspontaneous abortion. None of the women had undergone orwas given any treatment, such as cortisone, P, heparin, aspirin,or IV immunoglobulins (IVIg) or was immunized with allo-genic mononuclear cells at least 2 months before the tests.

The immunophenotypic tests involving the control group(n � 11) were carried out 6–18 months after their last livebirth (not breastfeeding). A written consent was obtainedfrom all women following the guidelines of the Greek Bio-ethical Committee.

Sample DescriptionIn total, 99 partner-specific women (88 subjects and 11

control group patients) were included in the study, whichwas comprised of 25 consecutive aborters (group I), 30sporadic aborters (group II), 33 infertile women (group III),and 11 fertile women (control group). The mean age of thetotal sample was 33.6 � 5.3 years (Table 2).

Flow-Cytometric Analysis of the PeripheralBlood Lymphocytes

Heparinized peripheral blood was obtained from the womenunder study at random phases of their menstrual cycle. Whiteblood cells and lymphocyte concentrations were manually andautomatically investigated (Cell Dynn 3500, Abbott SystemAnalyser), whereas mononuclear cells were immediately sepa-rated with Ficoll-Hypaque gradient centrifugation, washedtwice with phosphate buffer saline (PBS), and resuspended atthe same buffer. For the characterization of the PBMCs on thebasis of their surface antigens, a two- or three-color immuno-fluorescence analysis was performed on an EPICS-PROFILE II(Coulter) flow cytometer.

The following monoclonal antibodies (mAbs) were used forstaining: anti-CD3 fluorescein isothiocyanate (FITC), CD19phycoerythrin (RD1), anti-CD4 FITC, anti-CD8 RD1, anti-CD56 RD1, anti-CD3 phycoerythrin-Texas Red (ECD) (Immu-notech, Marseilles, France), and anti-CD16 FITC (Coulter, Mi-ami, FL). Briefly, the labeling process included incubation of 50�L from the PBMC preparation with the mAbs for 90 minutesat 4°C, two washes, and final resuspension in PBS. For thegating procedure, staining with anti-CD45 FITC and anti-CD14PE (Immunotech) was performed to analyze the CD45�/CD14� population. Isotype-matched FITC, PE, and ECD con-

trols (Immunotech) were used to exclude nonspecific binding.For each sample, 104 lymphocytes were evaluated. The two-and three-color flow cytometric analysis was performed asdescribed elsewhere (15).

The endometrial-type NK cells of the peripheral blood(CD56�/CD16�/CD3�) had a CD56� brighter expression incontrast to CD56�/CD16�/CD3� and CD56�/CD56�/CD3� lymphocytes. The percentage of CD56�/CD16� inthe total peripheral blood (PB) NK cell ratio calculation isgiven by the measurement of the CD3� cells, which areestimated by gating on the CD56PE/CD16 FITC primarylymphocyte analysis distribution.

Statistical AnalysisUnivariate analysis, consisting of general linear models

(GLMs) and one-way analysis of variance (ANOVA) wasused to test for intergroup differences in the following pa-rameters: total NK cell concentration (CD56�/CD16�/CD3�, CD56�/CD16�/CD3� and CD56�/CD16�/CD3�),percentage of total NK cells as a fraction of lymphocytes (%CD56�/CD16�/CD3�, CD56�/CD16�/CD3�, and CD56�/CD16�/CD3), endometrial-type NK cell concentration(CD56�/CD16�/CD3�), percentage of endometrial-typeNK cells as a fraction of the total NK cell population(%/TOTAL

ENDOMETRIAL TYPE NK), T-helper cell concentration(CD3�CD4�), T-cytotoxic cell concentration (CD3�CD8�),and B-cell concentration (CD19�).

To eliminate the impact of concomitants, patient age,white blood cell (WBC) count, lymphocyte concentration,and percentage of lymphocytes as a fraction of the WBCpopulation were treated as independent covariables. TheSheffe technique served as a tool for detecting pairs of studygroups that accounted for statistically significant differences.

Multivariate analysis, based on the application of GLMand t-tests, was used to investigate the relationship betweenfertility status and the aforementioned parameters. For thispurpose, two categories of patients were formed and com-pared. The first consisted of fertile women, irrespective oftheir abortion history (groups I and IV), while the secondincluded all infertile subjects (groups II and III).

The putative associations between the concentrations ofNK cells and lymphocytes, of endometrial-type and total NK

T A B L E 2

Mean age of the sample, base:total sample (99 subjects).

GROUP

Mean1. Consecutive

aborters2. Sporadic

aborters 3. Infertile 4. Fertile

Age (years) Mean 30.4 34.7 35.4 32.5 33.6SD 4.1 5.3 5.4 3.9 5.3


FERTILITY & STERILITY� 693

cells, and of CD3�/CD4� and B-cells were examined by theapplication of Pearson’s test to each study group separately.

RESULTSA summary of the results obtained by flow cytometric

analysis of PB samples is presented in Table 3.

PB NK Cells

Endometrial-Type NK Cells

Comparisons concerning the fraction. The percentageof endometrial-type PB NK cells in total PB NK cells(%/TOTAL PB

ENDOMETRIAL TYPE NK cells) indicates that fertile sub-jects (groups I and IV) score higher than infertile subjects(groups II and III) (Fig. 1). A statistically significant differ-ence was detected between study groups (groups I and IV vs.groups II and III) (PGLM of F-statistic �.001, PANOVA ofF-statistic �.001). In addition, the Sheffe technique traced astatistically significant difference between the followinggroups: between group I and group II (P�.002), betweengroup I and group III (P�.009), between group IV and groupII (P�.001), and between group IV and group III (P�.003).

The results mentioned above were in line with the resultsobtained by the application of the GLM and the t-test inwhich the fertility status was an independent factor. Morespecifically, both tests showed a significant difference be-tween the fertile (groups I and IV) and the infertile (groupsII and III) groups (P�.001 in both tests).

Comparison concerning the CD56�/CD16�/CD3� con-centration. Regarding the endometrial-type NK (CD56�/

CD16�/CD3�) cell concentration, a significant correlationwas revealed when study groups (all groups) (PGLM � .002)and fertility status (groups I and IV vs. groups II and III)(PGLM�.001) were analyzed, although none of these rela-tions could be confirmed by one-way ANOVA or t-test,respectively. According to Pearson’s test, endometrial-typeand total NK cell concentrations were significantly corre-lated in groups I (r � .681, P�.001), II (r � .708, P�.001),and III (r � .489, P�.004), but not within the group of fertilecontrols (group IV).

When the aforementioned tests were repeated in randomphases of the women’s menstrual cycles, the percentage ofendometrial-type NK in total NK cells remained the same.

Total PB NK Cells

Multivariate analysis revealed that the percentage of totalPB NK cells as a fraction of lymphocytes was affected byfertility status, with sporadic aborters and infertile subjectsexhibiting significantly higher values than fertile ones(PGLM�.004). Additionally, significant differences were de-tected among the studied groups (PGLM�.018 andPANOVA�.003). Higher values were traced in sporadic abort-ers (group II) than in control subjects (PSheffe�.019).

Total NK cell counts were found to be higher in sporadicaborters and infertile (groups II and III) than in fertile women(groups I and IV) (PGLM�.017). Statistically significant differ-ences in total NK cell concentration were found among the fourstudy groups (PGLM�.032 and PANOVA�.004). Total NK con-centration was increased in the subjects of groups I, II, and IIIwith respect to the control group (group IV). This increase was

T A B L E 3

Summary of results obtained by flow cytometry.

Flow-cytometric parameters

Patients classified by

Fertility/abortion history Fertility status

Consecutiveaborters

(group 1)

Sporadicaborters

(group 2)Infertile

(group 3)

Fertilecontrols

(group 4)Fertile

(groups 1, 4)Infertile

(groups 2, 3)

Total PB NK cell concentration (n/�L)

217 � 138.8a 299.8 � 173.5 272.2 � 115.7 131 � 63.1 190.8 � 126.4 285.3 � 145.5

% Endometrial-type NK cells onlymphocytes

7.8 � 4.2 11.4 � 5.7 10.3 � 4.7 6 � 3.1 7.3 � 3.9 10.8 � 5.2

Endometrial-type NK cellconcentration (n/�L)

16.4 � 9.9 11.5 � 7.7 12.7 � 9.4 10.7 � 4.7 14.6 � 9 12.1 � 8.6

The fraction of % endometrial-typeNK cells/total PB NK cells

8.4 � 4.5 4.3 � 2.4 4.9 � 3.1 10 � 6.5 8.9 � 5.1 4.7 � 2.8

Helper T-cell concentration (n/�L) 1459.4 � 374.4 1413.4 � 397.4 1463.7 � 467.4 1173 � 211.4 1377.5 � 357.8 1439.7 � 432.7Cytotoxic T-cell concentration (n/

�L)702 � 250.4 580.2 � 208.3 680.2 � 246.9 596.8 � 179.2 672 � 234.7 632.6 � 233

B-cell concentration (n/�L) 252.4 � 109.3 278.6 � 145.3 333 � 206.4 200.4 � 85 236.5 � 104.1 307.1 � 180.6

Note: NK � natural killer; PB � peripheral blood.a Numerical values represent sample means � SDs.



statistically significant in sporadic aborters (group II) and infer-tile (group III) groups (PSheffe�.01 and .04, respectively). Astatistically significant correlation between the concentration ofNK cells and that of lymphocytes was detected only in group II(Pearson’s test, r � .429, P�.018).

WBC and the Rest of the LymphocyticPopulations

White blood cell (WBC) counts and lymphocyte concen-trations were higher in the infertile groups (groups II and III)

than in the fertile ones (groups I and IV) (Table 4). However,this difference (PGLM of F-statistic�.05) was not significanteither among the four study groups or between fertile andinfertile subjects.

In contrast to the aforementioned variables, CD3�/CD4�

T-cell concentrations appeared to be increased in all studygroups compared with the control subjects (Table 3) andwere not shown to be statistically associated with the studygroup or fertility status (P �.05). Similarly with respect to

F I G U R E 1

Mean (�SD) percentage of endometrial-type NK cells in total peripheral blood NK cells per study group. Infertile groups 2 and3: 4.7 � 2.3; and fertile groups 1 and 4: 8.9 � 5.1, base:total sample (99 subjects).


T A B L E 4

Mean values of white blood cell (WBC) and lymphocyte concentrations (base:total sample, 99 subjects)

Study group Fertility status

Consecutiveaborters

Sporadicaborters Infertile Fertile Infertile Fertile

WBC concentration (n/�L) Mean 7,460.0 7,850.0 8,155.2 7,490.9 8,009.8 7,469.4SD 2,048.8 2,218.4 1,892.6 2,085.9 2,043.1 2,030.2

Lymphocyte concentration (n/�L) Mean 2,545.7 2,615.0 2,799.1 2,256.8 2,711.5 2,457.4SD 678.4 703.7 728.7 496.9 717.1 635.9

Lymphocyte % in WBCs Mean 34.7 34.1 35.0 31.2 34.5 33.6SD 6.8 6.8 8.4 7.4 7.6 7.1



the CD3�/CD8� T and B lymphocyte concentrations, nosignificant differences among the four study groups weredetected. Nevertheless, CD3�/CD4� and B-cell concentra-tions were related in the groups of sporadic aborters andinfertile subjects (r � .420, P�.021 for sporadic aborters; r� .672, P�.001 for infertile patients).

Age Distribution of Study GroupsThe differences shown in the Table 2 among the study

groups is of no clinical significance, but the one-wayANOVA technique indicated a statistically significant dif-ferentiation among the four study groups (P of F-statistic �.001). The Sheffe technique traced a statistically significantdifference between group I and group II (P�.019) andbetween group I and group III (P�.003). Thus, the variableof age is used as an independent covariate in all GLMsapplied afterwards.

DISCUSSIONIt is well known that elevated levels of PB NK lympho-

cytes are a feature of some women suffering from sponta-neous abortions. The decisive factor for identification ofimmunopathological conditions for NK cell numbers is thattheir percentage is greater than 12% of the total PB lympho-cyte level (16). In the present study, 30% of infertile (groupIII) and 16% of fertile and sporadic aborters (groups I and II)met the previous criteria. On the basis of the one-wayANOVA and GLM tests, only groups II and III (P�.01 andP�.04, respectively), but not group I, were characterized bya statistically significant elevation of the total PB NK con-centration. Previous studies (6, 7, 16) indicate a statisticallysignificant rise of the PB NK proportion on total PB lym-phocytes in recurrent aborters, which is in partial agreementwith our results, since these studies do not differentiate thecases according to infertility status but only take into accountthe number of abortions.

The increase in the total PB NK concentration possiblyreflects the increased cytotoxic activity of the infiltrate ofNK cells in the decidua, which creates a hostile immuneenvironment. The invading trophoblast cells are character-ized by a diminished expression of HLA class I moleculesmaking them more susceptible to attack. This hypothesis isin agreement with Kwak et al. (17), where increased CD57�

NK cells were detected at the implantation site of recurrentspontaneous fertile aborters. Since the CD56�/CD16�/CD3� NK cells, which naturally colonize the late secretoryendometrium, do not express CD57� molecules (18), theelevation of total PB NK levels along with CD4� T lym-phocytes (although of no statistical significance) suggests animmunoreactive condition of unknown etiology. In thisstudy, the augmentation of the total PB NK cell concentra-tion observed in group I (217 � 138.8) is of no statisticalimportance when compared with the control group (131 �63.1). This can be attributed to the fact that abortion is notonly related to abnormal NK numbers, but to many other

factors such as infections, autoantibodies, chromosomal ab-normalities, and so on (19).

In the present study, aborters experiencing unproblematicconception (group I) were characterized by a proportionalelevation of all PB NK subpopulations, including endome-trial-type PB NK cells. As a result, the fraction of endome-trial-type PB NK cells to total PB NK cells remained almostinvariable compared with the control group.

In infertile (group III) compared with fertile controls(group IV), the elevation of total PB NK cells (P�.04) wasaccompanied by a decrease in the ratio of endometrial-typePB NK cells to total PB NK cells (P�.003). Furthermore,when the factor infertility was evaluated between aborters(groups I and II), the concentration of total PB NK cells wasnot significantly increased in contrast to the ratio of endo-metrial-type NK to total PB NK cells, which was diminished(P�.002). The results concerning the above fraction suggestthat this parameter may be important for the implantation,since in this study a large number, from both groups II andIII, of the subjects had undergone IVF treatment with formedembryos.

The fraction, %/TOTAL PBENDOMETRIAL TYPE NK cells in effect rep-

resents the supply from the maternal circulation of the cellsfirst described as K-cells. Recent findings in recurrent spon-taneous abortion cases indicate a decidual reduction ofCD56�/CD16�/CD3� (20) in contrast to the cytotoxicCD16�. The fraction’s value is directly affected by thedenominator’s elevation, which in turn reflects the total PBNK concentration. Autoimmune diseases, although impli-cated in recurrent pregnancy loss (21), are not accompaniedby elevated NK levels. Infectious diseases, not necessarilylocalized in the endometrium, can contribute to increasedNK cell levels, taking into account that immune responses toinfections involve Th1 cytokine production. Activity of LAKcells becomes apparent only after the action of IL-2 on thePB NK cells, and a characteristic of aborters is that their IL-2levels are elevated (22). This could partially explain thereduction observed in the ratio of CD56�/CD16�/CD3� PBNK cells (endometrial-type NK cells) to the total PB NKlevels in infertility cases (groups II and III; Table 3).

The mean numbers of the endometrial-type NK cellsratios or concentration were nearly the same in both groupsII and III. In both groups, eight out of 11 infertile women(72.7%) with normal total NK concentrations, have a re-duced ratio of endometrial-type NK to total PB NK cells(�5.5%). In contrast, 66% of the women with a reducedratio of endometrial-type NK to total PB NK cells had anincreased total NK concentration. These observations indi-cate that the diminished fraction is not only dependent on theelevation of the total NK lymphocyte concentration.

Although CD56�/CD16�/CD3� NK cells are the mainlymphocytic population of the endometrium in the late se-cretory phase, reaching up to 70%–90% of the total periph-


eral lymphocytic population, they represent a very low, butconstant throughout the menstrual cycle, proportion in thePB. This observation indicates that hematopoiesis suppliesthese cells to the blood, hence sustaining a constant propor-tion, after their migration to the secretory endometrium.

CD56�/CD16�/CD3� cells secrete various cytokines, in-cluding macrophage colony-stimulating factor (M-CSF) andgranulocyte M-CSF (MG-CSF), which are capable of promot-ing placental growth (23). Their presence in the decidua andthe endometrium has been shown (K-cells), but their func-tion still remains largely unknown. Although studies haveshown that they contribute to the shedding of the nonim-planted endometrium, through relaxin production, no suchrole has been observed in the decidua. It is possible thatCD56�/CD16�/CD3� cells can have a suppressive role inthe endometrium through local production of immunosup-pressive factors without exhibiting increased cytotoxic ac-tivity (24). The mechanism of their blood concentrationmaintenance and of their supply in the endometrium has notbeen yet elucidated. It is possible that Th1 cytokines, whichprevail in aborters and infertile women (25), cause the ele-vation of total NK cells while having no similar effect onendometrial-type NK cells.

Fukui et al. (14) studied the percentage of the CD56� inendometrial samples of infertile patients undergoing IVF byobtaining endometrial tissue at the midsecretory phase oftheir menstrual cycle before IVF treatment . They found nodifference in the proportion of CD56�/CD16� cells betweenthe failed and implanted group, but taking into account thatthe endometrial-type NK cell concentration only reaches amaximum in the late secretory phase, the result was probablydue to the early tissue collection. In the present study, themean numbers of the PB ratio of those cells (CD56�/CD16�/CD3�) to the total NK cells was found to be low inwomen experiencing infertility (4.7 � 2.8). On the otherhand, the mean numbers of the fertile women were high (8.9� 5.1), irrespective of their pregnancy outcome.

It seems that these cells contribute to the establishment ofa suitable implantation environment, in contrast to the rest ofthe NK subpopulations, which favor the establishment of ahostile one. These observations correspond to preliminarydata showing that PB CD56�/CD16�/CD3� cells could beused as an indicator of subsequent successful implantationand maintenance of the gestation. Therefore, this fraction ofPB NK cells could be a good prognostic feature, especiallybefore any infertility-related treatment. Further investigationinvolving a prospective study is required to verify theseobservations.

Acknowledgments: The authors thank Mr. P. Katerelos for his help with thestatistical analysis and Miss K. Andriopoulou for her technical assistance.

References1. Lopez-Botet M, Moretta L, Strominger J. NK-cells receptors and HLA

class I molecules. Immunol Today 1996;212–4.2. Moretta A, Biassoni R, Bottino C, Mingari MC, Moretta L. Natural

cytotoxicity receptors that trigger human NK-cell–mediated cytolysis.Immunol Today 2000;21:228–34.

3. Nagler A, Lanier LL, Cwirla S, Philips JH. Comparative studies ofhuman FcRIII-positive and negative natural killer cells. J Immunol1989;143:3183–91.

4. King A, Hiby SE, Verma S, Burrows T, Gardner L, Loke YW. UterineNK cells and trophoblast HLA class I molecules. Am J Reprod Immu-nol 1997;37:459–62.

5. Aoki K, Kajiura S, Matsumoto Y, Ogasawara M, Okada S, Yagami Y,et al. Preconceptional natural killer cell activity as predictor of miscar-riage. Lancet 1995;345:1340–2.

6. Coulam CB, Coodman C, Roussev RG, Thomason EJ, Beaman KD.Systemic CD56�cell can predict pregnancy outcome. Am J ReprodImmunol 1995;34:93–9.

7. Yamada H, Kato EH, Kobashi G, Ebina Y, Shimada S, Morikawa M,et al. High NK cell activity in early pregnancy correlates with subse-quent abortion with normal chromosomes in women with recurrentabortion. Am J Reprod Immunol 2001;46:132–6.

8. Ruis JE, Baum L, Gilman-Sachs A, Beaman KD, Kim Y, Beer AC.Intravenous immunoglobulin inhibits natural killer cell activity in vivoin women with recurrent spontaneous abortions. Am J Reprod Immunol1996;35:370–5.

9. Ramhorst R, Argiello E, Zittermann S, Pando M, Larriba J, Irigoyen M,et al. Is the paternal mononuclear cells’ immunization a successfultreatment for recurrent spontaneous abortion? Am J Reprod Immunol2000;44:129–35.

10. Jenkins C, Roberts J, Wilson R, MacLean MA, Shilito J, Walkers JJ.Evidence of a TH 1 type response associated with recurrent miscar-riage. Fertil Steril 2000;73:1206–8.

11. Szereday L, Vagra P, Szekeres-Bartho J. Cytokine production by lym-phocytes in pregnancy. Am J Reprod Immunol 1997;38:418–22.

12. King A, Kalra P, Loke YW. Human trophoblast cell resistance todecidual NK lysis is due to lack of NK target structure. Cell Immunol1990;127:230–7.

13. Clark DA, Arck PC, Chaouat G. Why did your mother reject you? Im-munogenetic determinants of the response to environmental selective pressureexpressed at the uterine level. Am J Reprod Immunol 1999;41:5–22.

14. Fukui A, Fujii S, Yamaguchi E, Kimura H, Sato S, Saito Y. Naturalkiller cell subpopulations and cytoxicity for infertile patients undergo-ing in vitro fertilization. Am J Reprod Immunol 1999;41:413–22.

15. Gruber R, Reiter C, Riethmuller G. Triple immunofluorescence flowcytometry, using whole blood, of CD4� and CD8� lymphocytesexpressing CD45RO and CD45RA. J Immunol Meth 1993;163:173–9.

16. Beer AE, Joanne YH, Ruiz JE. Immunophenotypic profiles of periph-eral blood lymphocytes in women with recurrent pregnancy losses andin infertile women with multiple failed in vitro fertilization cycles.Am J Reprod Immunol 1996;35:376–82.

17. Kwak JYH, Beer AE, Kim SH, Mantouvalos HP. Immunopathology ofthe implantation site utilizing monoclonal antibodies to natural killercells in women with recurrent pregnancy losses. Am J Reprod Immunol1999;41:91–8.

18. Chernyshov VP, Slukvin II, Bondarenko GI. Phenotypic characteriza-tion of CD7�, CD3�, and CD8� lymphocytes from first trimesterdecidua using two-color flow cytometry. Am J Reprod Immunol 1993;29:5–16.

19. Yamamoto T, Takahashi Y, Kase N, Mori H. Role of decidual naturalkiller (NK) cells in patients with missed abortion: differences betweencases with normal and abnormal chromosome. Clin Exp Immunol1999;116:449–56.

20. Yamamoto T, Takahashi Y, Kase N, Mori H. Decidual natural killercells in recurrent spontaneous abortion with normal chromosomal con-tent. Am J Reprod Immunol 1999;41:337–42.

21. Wallace DJ, Druzin ML, Lahita RG. Clinical rheumatologic applica-tions of reproductive immunology. Facts, fiction, and fancy. ArthrRheum 1997;40:209–16.

22. Hamai Y, Fujii T, Kozuma S, Shibata Y, Taketani Y. Secretion ofinterleukin-2 from unstimulated peripheral blood mononuclear cells isa possible pathogenic mechanism in recurrent abortion. Am J ReprodImmunol 1998;40:63–4.

23. Saito S, Nishikawa K, Morii T, Enomoto M, Narita N, Motoyoshi K, etal. Cytokine production by CD16-CD56 bright natural killer cells in thehuman early pregnancy decidua. Int Immunol 1993;5:559–63.

24. Nagler A, Lanier LL, Cwirla S, Phillips JH. Comparative studies ofhuman FcRIII-positive and negative natural killer cells. J Immunol1989;143:3183–91.

25. Lim KJH, Odukoya OA, Ajjan RA, Li TC, Weetman AP, Cooke ID.The role of T-helper cytokines in human reproduction. Fertil Steril2000;73:136–42.


1-s2.0-S0015028203007787-main

Documents

Transcript of 1-s2.0-S0015028203007787-main