Cutting Edge: Heterozygote Advantage in Autoimmune
Disease: Hierarchy of Protection/Susceptibility
Conferred by HLA and Killer Ig-Like Receptor
Combinations in Psoriatic Arthritis1
George W. Nelson,2* Maureen P. Martin,2* Dafna Gladman,†Judith Wade,‡
John Trowsdale,§and Mary Carrington3*
like receptor (KIR) genotypes influence resistance to sev-
studies is challenging because it involves multiple linked
and unlinked loci that exert their influence in an epistatic
activating receptors were susceptible to developing psori-
were missing. In this study, we present a novel model in
which susceptibility to PsA is determined by the overall
for increasing susceptibility to PsA with more activating
of KIR/HLA combinations on this disease. The Journal
of Immunology, 2004, 173: 4273–4276.
mortality generates and maintains diversity at this locus (1–3).
Although haplotypes containing multiple activating KIR may
mediate protective NK cell responses against infectious disease
(2) and enhance beneficial maternal-fetal interactions,5these
same haplotypes may predispose to autoimmune pathogenesis.
Indeed, activating KIR, in particular KIR2DS1 and KIR2DS2,
have been associated with autoimmune disorders (2, 3).
KIR2DS1 and KIR2DS2 show 98% sequence similarity to
everal human diseases have been shown to associate
with variability at the killer Ig-like receptor (KIR)4gene
cluster, supporting a model whereby disease morbidity/
of the same gene and their products recognize the same HLA
ligands), respectively, in their extracellular domains, but the in-
KIR2DL1 and HLA-Cw group 1 for KIR2DL2/3) with signif-
icantly greater affinity than do the activating KIR2DS (4). De-
coding KIR2DL1 and KIR2DL2/3 are present in virtually all
individuals, unlike their activating counterparts, KIR2DS1 and
KIR2DS2, which are present in ?35% and 56% of European
effect was exacerbated in the absence of ligands for the corre-
sponding inhibitory NK cell receptors, KIR2DL1 and
KIR2DL2/3, respectively (odds ratio ? 3.2, 95% confidence
interval ? 1.8–5.5, p ? 3 ? 10?5), compared with subjects
lacking both KIR2DS1 and KIR2DS2). For example, individu-
the HLA-Cw group 2 ligands for KIR2DL1 were absent; the
same was true of KIR2DS2 when HLA-Cw group 1 ligands for
KIR2DL2/3 were absent. We concluded that the presence of an
activating KIR was particularly detrimental when ligand for the
corresponding inhibitory KIR was missing, in which case an in-
hibitory signal would not quench a potentially harmful activat-
We reconsidered the model tested in this study because the-
be able to provide counteracting inhibition. The function of
NK cells is regulated by positive and negative signals transmit-
vivo, NK cells are under the constitutive dominance of inhibi-
*Basic Research Program, Science Applications International Corporation-Frederick, Na-
tional Cancer Institute, Frederick, MD 21702;†Toronto Western Research Institute and
Centre for Prognosis Studies in the Rheumatic Diseases, Toronto Western Hospital,
and‡Regional HLA Laboratory, University Health Network, Toronto, Ontario, Canada;
and§Immunology Division, Department of Pathology, University of Cambridge, Cam-
bridge, United Kingdom
Received for publication June 22, 2004. Accepted for publication August 3, 2004.
This article must therefore be hereby marked advertisement in accordance with 18 U.S.C.
Section 1734 solely to indicate this fact.
Institutes of Health, under Contract NO1-CO-12400.
2G.W.N. and M.P.M contributed equally to this work.
3Address correspondence and reprint requests to Dr. Mary Carrington, National Cancer
Institute-Frederick,P.O. Box B,Frederick,
4Abbreviations used in this paper: KIR, killer Ig-like receptor; PsA, psoriatic arthritis.
5S. E. Hiby, J. J. Walker, K. M. O’Shaughnessy, C. W. G. Redman, M. Carrington, J.
Trowsdale, and A. Moffett. Combinations of maternal KIR and fetal HLA-C genes influ-
ence the risk of pre-eclampsia and reproductive success. Submitted for publication.
Copyright © 2004 by The American Association of Immunologists, Inc.0022-1767/04/$02.00
functions occur only when activating signals are able to over-
come inhibitory signaling. This is achieved either by a predom-
inance of activating receptor:ligand interactions or a lack of in-
hibitory receptor:ligand interactions (9). KIR are clonally
cell clone expresses only a portion of the genes within the ge-
netic profile (8, 10). This predicts that, depending on the ge-
notype, a given individual could have a relatively high fre-
quency of NK cells that are: 1) primarily under the control of
inhibitory receptors (most inhibition); 2) controlled by inhibi-
tory and activating receptors fairly equally (relatively neutral);
or 3) primarily under the control of activating receptors (most
activation). In a similar vein, individuals who are missing li-
gands for some inhibitory receptors (as is the case among
HLA-Cw group 1 or 2 homozygotes) will have fewer NK cells
under inhibitory control than individuals who have all ligands
present. Thus, an activating KIR, such as KIR2DS1, might be
detrimental in terms of developing PsA if the ligand for either
KIR2DL1 or KIR2DL2/3 is missing (i.e., homozygotes for ei-
ther group of HLA-Cw ligands). Based on this reasoning, we
of KIR expression and function, and that shows more robust
statistical support for the role of KIR in susceptibility to PsA
than does our previous model.
Materials and Methods
The study population and methodology are as previously described (2). The
trends for increasing susceptibility to PsA for combined KIR-HLA genotypes,
under the two models, were evaluated by the Mantel-Haenzel test (SAS PROC
FREQ; SAS Institute, Cary, NC). Analysis comparing the effect of multiple
factors on PsA used logistic regression (SAS PROC LOGISTIC; Cary Insti-
Results and Discussion
We recently reported a strong association between KIR geno-
(2). Based on the model used in the analyses, we proposed that
KIR2DS1 and KIR2DS2 (both activating receptors with un-
known high affinity ligands) increase the risk of developing
PsA, particularly when HLA class I ligands for the correspond-
ing inhibitory KIR (KIR2DL1 and KIR2DL2/3, respectively)
troduction), we now hypothesize that susceptibility to PsA in-
creases progressively with increasing levels of KIR-mediated ac-
presence of certain activating KIR and with the absence of li-
gand for inhibitory KIR. The absence of ligand for inhibitory
KIR occurs when individuals are homozygous for either
for KIR2DL1). Therefore, we tested for effects of HLA-Cw
four groups according to the presence or absence of the activating KIR2DS1 and/or KIR2DS2, and zygosity (hom, homozygosity; het, heterozygosity) for the
most susceptible (top) to most protected (bottom). The two middle groups were combined in the test for trend, as both represent intermediate susceptibility. B, Test
susceptible under the new model (red) are compared with subjects predicted to be susceptible under both models (yellow) and subjects predicted to be intermediate
in terms of risk under both models (blue).
A, Trend toward increasing susceptibility to PsA with genotypes conferring increasing KIR-mediated NK cell activation. Subjects are divided into
4274 CUTTING EDGE: HETEROZYGOTE ADVANTAGE IN AUTOIMMUNE DISEASE
group zygosity in combination with the presence/absence of
KIR2DS genes based on genotypes expected to result in a range
of NK cell activity from the most activating to the most inhib-
1 or 2 homozygous; 2) KIR2DS1 and/or KIR2DS2, HLA-Cw
group heterozygous; 3) no KIR2DS1 or KIR2DS2, HLA-Cw
and 3, as there is no a priori reason why one should be more
susceptible than the other); and 4) no KIR2DS1 or KIR2DS2,
HLA-Cw group heterozygous (Fig. 1A). A ?2test of 26.8 for
trend (p ? 2 ? 10?7) was determined using this new model
an activating KIR in the absence of ligand for corresponding
Statistical support for the old model, although weaker than
that for the new model, was still quite strong, probably because
lap somewhat. In fact, all individuals in the most susceptible
group under the old model are also in the most susceptible
group under the new model because an individual who lacks a
characteristics of the most susceptible group in the old model)
are necessarily homozygous for HLA-Cw group, because Cw
group heterozygotes necessarily carry the ligand for both inhib-
itory receptors. In contrast, some individuals in the most sus-
ceptible group under the new model were not included in the
the old model, some individuals who were homozygous for
HLA-Cw group were classified in an intermediate group. These
included: 1) those who had KIR2DS1 in the absence of
for the corresponding KIR2DL1); and 2) those who had
Cw group 1 (i.e., the ligand for the corresponding KIR2DL2/
3). These individuals are classified under the most susceptible
group and positive for KIR2DS1 and/or KIR2DS2. Thus, the
level of susceptibility of these particular genotypes allows direct
determination of the model that best fits the genotypic data. As
shown in Fig. 1B, individuals who are homozygous for Cw
group are at increased risk of developing PsA regardless of
whether ligand for corresponding inhibitory KIR is present or
absent (compare the top three sets of bars, where yellow repre-
and red combined represent individuals susceptible in the new
model). All three sets of homozygous individuals show very
similar levels of susceptibility to one another, but they are dis-
tinct from groups who are heterozygous for Cw group (blue
bars). These data fit the prediction of the new but not the old
The primary difference between the old and new models is
Table I. Influence of KIR and HLA on psoriatic arthritis
KIR2DS1 and/or KIR2DS2
HLA-Cw group homozygosity
KIR2DS1 and/or KIR2DS2, without ligand
ence of another inhibitory receptor-ligand combination (in this case, KIR2DL2/3 ?
most activation (susceptibility). A putative protective genotype, 2DL1-2DL2/3; Cw
with heterozygosity for HLA-Cw group is shown in the top panel (pink). The middle
group 1, in which one or both activating KIR2DS are present in combination with
HLA-Cw group heterozygosity; and ii, 2DL1-2DL2/3, Cw group 2 in which no acti-
vating KIR2DS are present in combination with HLA-Cw group homozygosity (yel-
4275The Journal of Immunology
inhibitoryreceptors.Intheoldmodel(Fig.2A),theassumption Download full-text
was made that only a corresponding inhibitory receptor could
quench the activity of an activating receptor. Therefore, activa-
KIR2DL1, but not by KIR2DL2 (Fig. 2A). The new model,
in contrast, does not make that assumption; rather, it proposes
that the trend for susceptibility to develop PsA increases
when genotypes are ordered by their ability to confer the most
inhibition (protection) to the most activation (susceptibility)
We further tested five explanatory factors (HLA-B*27, HLA-
Cw*0602, presence of KIR2DS1 and/or KIR2DS2, presence of
KIR2DS1 and/or KIR2DS2 in the absence of HLA-Cw group
ligand for the corresponding inhibitory KIR, and HLA-Cw
group homozygosity) in a multiple logistic regression (Table I).
The presence of KIR2DS1 and/or KIR2DS2 is highly signifi-
cantly associated with susceptibility to PsA, as are the estab-
lished factors HLA-B*27 and HLA-Cw*0602 (p ? 0.0001).
HLA-Cw group homozygosity is also significantly associated
(p ? 0.80) in the multiple regression analysis. We also consid-
ered the possibility that, like heterozygosity for HLA-Cw group
1 and 2, HLA-B Bw4, the ligand for KIR3DL1, may provide
ever, multiple logistic regression analysis indicated no protec-
tive effect of HLA-B Bw4 among homozygotes for Cw group,
heterozygotes for Cw group, or all subjects combined. Thus,
protective inhibitory effects against PsA appear to be restricted
standing of KIR expression and function, a model proposing
that a combination of KIR2DS1 and/or KIR2DS2 plus
HLA-Cw ligand group homozygosity (a situation where the in-
hibitory signal is diminished) confers susceptibility to PsA is
It is essential to continually pursue plausible models for data
pertaining to complex genetic loci (such as HLA and KIR) in
human diseases. If logical intelligent models are used in data
ing future biological studies. One general model is unlikely to
fit data derived from different types of diseases, and as our un-
derstanding of KIR biology advances, we must reconsider and
strengthen and clarify the genetic effects of these polymorphic
loci on human disease.
1. Martin, M. P., X. Gao, J. H. Lee, G. W. Nelson, R. Detels, J. J. Goedert,
S. Buchbinder, K. Hoots, D. Vlahov, J. Trowsdale, et al. 2002. Epistatic interaction
between KIR3DS1 and HLA-B delays the progression to AIDS. Nat. Genet. 31:429.
2. Martin, M. P., G. Nelson, J. H. Lee, F. Pellett, X. Gao, J. Wade, M. J. Wilson,
J. Trowsdale, D. Gladman, and M. Carrington. 2002. Cutting edge: susceptibility to
psoriatic arthritis: influence of activating killer Ig-like receptor genes in the absence of
specific HLA-C alleles. J. Immunol. 169:2818.
3. Yen, J. H., B. E. Moore, T. Nakajima, D. Scholl, D. J. Schaid, C. M. Weyand, and
J. J. Goronzy. 2001. Major histocompatibility complex class I-recognizing receptors
are disease risk genes in rheumatoid arthritis. J. Exp. Med. 193:1159.
4. Vales-Gomez, M., H. T. Reyburn, R. A. Erskine, and J. Strominger. 1998. Differen-
tial binding to HLA-C of p50-activating and p58-inhibitory natural killer cell recep-
tors. Proc. Natl. Acad. Sci. USA 95:14326.
5. Tomasello, E., M. Blery, F. Vely, and E. Vivier. 2000. Signaling pathways engaged by
NK cell receptors: double concerto for activating receptors, inhibitory receptors and
NK cells. Semin. Immunol. 12:139.
6. Moretta, A., R. Biassoni, C. Bottino, and L. Moretta. 2000. Surface receptors deliv-
ering opposite signals regulate the function of human NK cells. Semin. Immunol.
and NK cell recognition. Immunol. Today 11:237.
D’Andrea, J. H. Phillips, L. L. Lanier, and P. Parham. 1997. Functionally and struc-
turally distinct NK cell receptor repertoires in the peripheral blood of two human
donors. Immunity 7:739.
9. Lanier, L. L. 2001. Face off–the interplay between activating and inhibitory immune
receptors. Curr. Opin. Immunol. 13:326.
10. Shilling, H. G., N. Young, L. A. Guethlein, N. W. Cheng, C. M. Gardiner, D. Tyan,
and P. Parham. 2002. Genetic control of human NK cell repertoire. J. Immunol.
4276CUTTING EDGE: HETEROZYGOTE ADVANTAGE IN AUTOIMMUNE DISEASE