REJUVENATION RESEARCH

Volume 14, Number 5, 2011

ª Mary Ann Liebert, Inc.
DOI: 10.1089/rej.2011.1169

The Genetics of the Human APOE Polymorphism

Davide Seripa,1 Grazia D’Onofrio,1 Francesco Panza,1 Leandro Cascavilla,1

Carlo Masullo,2 and Alberto Pilotto1,3

Abstract

The genetic origin of the three common variants of the human apolipoprotein E (apoE) protein, known as E2, E3
and E4, was understood in 1981, and since the mid 1980s these are probably the most-studied protein variants in
human races. They have been related to a number of age-related diseases, including Alzheimer disease, as well
as to healthy aging and longevity. The gene variants underlying these protein isoforms, known as e2, e3, and e4,
are allelic forms of the APOE gene, resulting from different haplotypes at the APOE locus (19q13.31). In
particular, they result from three of the four haplotypes expected by the combinations of the alleles of the two
single-nucleotide polymorphisms rs429358 and rs7412. The fourth missing haplotype, known as e3r, has been
identified in only two Caucasian families from Italy and in one Yoruba family from Nigeria worldwide. Thus,
this fourth APOE gene variant is rare, and it encodes a protein isoform, identified as E3r, showing identical
physical characteristics to E3, that conversely, is the most common form of apoE in humans. In this review
article, we report the identification of the haplotype e3r in a third Caucasian family from Italy, and then attempt
to re-examine the current knowledge regarding the APOE polymorphism, taking into account this fourth
haplotype. We also focus on the commonly accepted hypothesis for the evolution of the common APOE gene
variants, in which we include the e3r haplotype, previously not considered.

Introduction ants. In contrast, the proteins of class a (a-II, a-III, and a-IV)
represent heterozygosity for three different APOE gene vari-

T he three common variants of the apolipoprotein E

(APOE) gene are probably the most prolifically pub-
lished gene variants in human genome. It has been estimated
to account for more than 95% of the total genetic variation
showed by the apoE protein in Caucasians.1 This protein
plays a pivotal role in lipid metabolism. It is essential for the
normal catabolism of triglyceride-rich lipoprotein, and these
variants have been associated with a number of age-related
diseases, including atherosclerosis, cardiovascular disease,
and Alzheimer disease (AD), as well as with healthy aging
and longevity.2,3,4
The common variants of apoE were discovered about 40
years ago8 and their presence was confirmed at the end of
1970s in several studies.6–12 In particular, it was reported that,
upon two-dimensional gel electrophoretic analysis, apoE from
human very-low density lipoprotein (VLDL) of different in-
dividuals appears in either one of two complex patterns,
designated as class a and b.13 In 1981, it was demonstrated
that the proteins belonging to class b (b-II, b-III, and b-IV)
represent homozygosity for three identical APOE gene vari-
ants. Therefore, the three major apoE isoforms were geneti-
cally determined, and it was suggested they be identified with
the Greek letter epsilon (e).14 A further study in 1982 proposed
a nomenclature for these protein isoforms, to be identified as
E2, E3 and E4.15 These important findings led several re-
searchers to the identification of the apoE-encoding gene,
which was recognized in 1982,16 localized on chromosome 19
at locus q13.31,17 and sequenced in 1985.18 Three years later,
the DNAs encoding the three different protein isoforms were
definitively sequenced, and named as e2, e3, and e4.19 The
important implications of this study are the demonstration
that these common gene variants were generated by the two
single-nucleotide polymorphisms (SNPs) rs429358 and rs7412
in exon 4 of the APOE gene (AF261279), and that the variants
e2, e3, and e4 are indeed different haplotypes of the APOE
gene, generated by the combination of the allele of the two
SNPs at the APOE locus. This notion remained unnoticed for
about 20 years, thus obscuring interesting questions.20
As recently reported,20 the three common variants e2, e3,
and e4 of the APOE gene are determined by three of the four

1
Geriatric Unit & Gerontology-Geriatrics Research Laboratory, Department of Medical Sciences, IRCCS ‘‘Casa Sollievo della Sofferenza,’’
San Giovanni Rotondo (FG), Italy.
2
Institute of Neurology, Catholic University School of Medicine, Rome, Italy.
3
Geriatrics Unit, S. Antonio Hospital, Padova, Italy.

491
492
haplotypes resulting from the combination of the alleles of
the two SNPs rs429358 (C3937?T) and rs7412 (C4075?T) at
the APOE locus. In particular, the haplotype T3937-T4075
identifies the e2 variant, the haplotype T3937-C4075 identifies
the e3 variant, and the haplotype C3937-C4075 identifies the e4
variant. The fourth haplotype, C3937-T4075, identifies the
fourth variant, named e3r.21 It must be noted that nucleotides
3,937 and 4,075 correspond to the first base of two CGC
codons, encoding arginine (Arg), and each C?T transition
changes the first base in these codons, resulting in TGC co-
dons encoding cysteine (Cys). Thus, the e2 haplotype (T3937-
T4075) encodes the E2 protein isoform (Cys112-Cys158), the e3
haplotype (T3937-C4075) encodes the E3 protein isoform
(Cys112-Arg158), and the e4 haplotype (C3937-C4075) encodes
the E4 protein isoform (Arg112-Arg158). Accordingly, the e3r
haplotype (C3,937-T4,075) encodes the protein isoform E3r
(Arg112-Cys158). This fourth haplotype is rare, being previ-
ously identified only in two Caucasian families from Italy and
in one Yoruba family from Ibadan (Nigeria) worldwide.20–22
Notably, the 1000 Genomes Project did not report a high
variability for the APOE locus, describing a total of 63 vari-
ations, including 19 nonsynonymous coding variations. In
particular, no variations were found at rs429358 and
rs7412.23
The e3r Families
Family 1
This was the first family reported in the literature in which
the haplotype e3r was identified (Fig. 1).21 It was an Italian
family of Caucasian race. In a study investigating the
transmission rate of the e2 haplotype in families with autistic
probands, the e3r haplotype was identified in a male autistic
child (III:1) and in his mother (II:2). The authors named this
haplotype e3r, because it codes for an E3r protein isoform,
characterized by the same electric charge of the E3 isoform,
but with a Cys and an Arg in reverse order (Arg112-Cys158 in
the E3r protein isoform vs. Cys112-Arg158 in the E3 protein
isoform), and submitted this new sequence to the National
Center for Biotechnology Information (NCBI) GenBank (ac-
cession number AY077451). These authors, however, did not
find this haplotype in the further screening of 234 autistic
FIG. 1. The first family (Caucasian from Italy) in which the
e3r haplotype was observed. The patient (III:1) inherited the
e3r haplotype from his mother (II:2).
SERIPA ET AL.
patients, 514 first-degree relatives, and 182 unaffected con-
trols, thus concluding that the e3r haplotype is not associated
with autism.
Family 2
This was the second family described in the literature in
which the e3r haplotype was identified (Fig. 2).22 Indeed, it
also was the unique non-Caucasian, and in particular non-
Italian family, showing the e3r haplotype. In the In-
dianapolis-Ibadan Dementia Project,24 a longitudinal, com-
munity-based study that seeks to identify risk factors for
dementia in elderly African Americans living in In-
dianapolis, Indiana, and in elderly Yorubas residing in Iba-
dan, Nigeria, these authors identified the e3r haplotype in a
healthy 70-year-old Yoruba female (II:2) and her son (III:1),
with the same genotype. Notably, even though the DNA
sequence identifying this haplotype was already officially
identified in GenBank, these authors arbitrarily decided to
name this haplotype as e1y. They assigned number 1 because
it would be the next haplotype counting backward, assum-
ing e4 as the ancestral haplotype, and y for Yoruba, to dif-
ferentiate from the other e1 haplotype already described.
However, according to the accepted nomenclature for the
APOE variants,15 the e1 variant (also known as the Weis-
graber variant) is the only one, and encodes the E1 isoform
(Asp127- Cys158; Table 3).
Family 3
This was the third family described in the literature in
which the e3r variant was observed (Fig. 3). Indeed, this
was also the second Caucasian family identified in Italy. An
extensive description of this family was already reported.20
Briefly, in the context of a study investigating the potential
role of the APOE polymorphism in various type of late-
onset dementias, the e3r haplotype was identified in a 76-
year-old male patient affected by motor neuron disease
(MND) (II:1). Because only one study that reported inves-
tigating apoE in this disease,25 it was concluded that these
data were not sufficient to suggest an involvement of the
e3r haplotype in MND.
FIG. 2. The second family (non-Caucasian from Nigeria) in
which the e3r haplotype was observed. The subject (II:2)
transmitted the e3r haplotype to her son (III:1). II:1 is carrier
of an e4 haplotype.
THE FOUR APOE ALLELES 493

Table 2. ApoB/E Receptor Competitive

Binding Activity of VLDL from Patient
and Kindred of Family 3

Binding activity (%)

E4/E3r- E2/E4- E3/E4- E3/E3-

125
I-LDL VLDL VLDL VLDL VLDL
(mg/mL) (II:1) (III:1) (III:2) (Control)

0.5 101 102 97 92  6

1 96 91 82 89  8
3 85 89 87 81  11
9 73 68 70 76  9
18 52 59 57 59  6
36 37 38 39 41  8

Mean values of three measures on E3/E3-VLDL were used as

controls.
FIG. 3. The third family (the second Caucasian from Italy) ApoB, Apolipoprotein B; VLDL, very-low-density lipoprotein;
in which the e3r haplotype was observed. The patient (II:1) LDL, low-density lipoprotein.
did not transmit the e3r haplotype to his offspring.

Notably, because apoE plays an important role in lipid
metabolism, being crucial for the normal catabolism of tri-
glyceride-rich lipoproteins, possible alterations in lipid me-
tabolism were also investigated in the subject and his
offspring by determining a complete serum lipid and apoli-
poprotein profile (Table 1) as well as the apoB/E receptor
binding activity of e3r-containing VLDL (Table 2). With re-
spect to the other genotypes, the e3r haplotype did not seem
to significantly modify serum lipid and apolipoprotein levels
as well as the binding rate of VLDL to the apoB/E receptor.
Family 4
Recently, we identified this family, the fourth family
showing the e3r haplotype and the third Caucasian family
identified in Italy (Fig. 4). In the context of a study investi-
gating genetic factors in healthy and nonhealthy aging in
southern Italy,3,26 we identified the e3r haplotype in a 80-
year-old male nondemented, cognitively intact subject and
in his daughter. Also, normal serum lipid and apolipoprotein
profiles were observed in this family.
are in reverse order in respect to E3 (Arg112-Cys158) (Fig. 5).
Therefore, these two protein isoforms show the same elec-
trical charge as well as the same physical characteristics.
Accordingly, they were indistinguishable by using protein-
based methods of analysis, such as isoelectric focusing or
western blot analysis, the up-to-date the gold standard for
the analysis and nomenclature of the APOE isoforms.15 This
behavior of the E3r isoform suggested three main consider-
ations. First of all, given the similarity between the E3 and
the E3r, the observed normal serum lipid and lipoprotein
profiles as well as a normal binding rate of E3r-containing
VLDL with the apoB/E receptor are fully justified. Second,
the early use of protein-based methods for the analysis of the
APOE gene variants may have greatly contributed to un-
derestimating the frequency of e3r. Finally, it is clear that this
gold standard was currently replaced by newer molecular
biology techniques, such as the polymerase chain reaction
(PCR)-based procedures.27 However, despite the great im-
portance of the APOE genotyping in clinical laboratory anal-
ysis, as well as in the genetics of Alzheimer disease, healthy

The E3r Protein Isoform

The protein variant E3r is characterized by a cysteine (Cys)
at codon 112 and by an arginine (Arg) at codon 158, which

Table 1. Serum Lipid and Apolipoprotein Levels

in Patient and Kindred of Family 3

II:1 III:1 III:2

(e3r/e4) (e2/e4) (e3/e4)

Total cholesterol 169 181 196

IDL cholesterol 46 58 73
LDL cholesterol 100 99 113
VLDL cholesterol 23 15 10
Tryglicerides 115 88 52
Apolipoprotein B 85 86 74
Apolipoprotein E 5.0 5.1 4.9
Lipoprotein (a) 18 17 18
FIG. 4. The fourth family (the third Caucasian from Italy)
Values are in mg/dl. in which the e3r haplotype was observed. The patient (II:1)
IDL, Intermediate-density lipoprotein; LDL, low-density lipopro- transmitted the e3r haplotype to his daughter. II:2 has an e3/
tein; VLDL, very-low-density lipoprotein. e4 genotype.
494
FIG. 5. Three-dimensional models of the four main apoE protein isoforms.
aging, and longevity,2,4,28 a consensus statement on identify-
ing a standard DNA-based method for the identification of
the common APOE gene variants has not currently reported
and may be strongly suggested. Table 3 provides an idea of
the confusing status in the nomenclature of the apoE isoforms.
It may be observed that the genotypes at codons 112 and 158
of all these isoforms are always referred to one of the four
alleles. In contrast, the phenotypes, as assessed by a protein-
based approach, and the associated trivial names are largely
confusing.
The Genetics of the E3r Protein Isoform
The origin of the haplotype e3r may be explained by three
principal mechanisms (Fig. 6). In considering the APOE locus
on a chromosome 19 pair at meiosis (pathway A), a recom-
bination within the locus in a subject with an e2/e4 genotype
will result in 50% of e2/e4 genotypes and in 50% of e3/e3r
genotypes in the offspring. However, this intragenic recom-
bination may be not common. In fact, the presence of a
strong linkage between rs429358 and rs7412 is well known.1
As a cause of this linkage, we observed linkage disequilib-
rium between these two SNPs in the analysis of 1,344 healthy
Caucasians from the same genetic background as families 3
and 4 (D0 > 0.500).20 The second possibility (pathway B) is a
mutation occurring in an e2 haplotype at nucleotide 3,937,
changing a T to a C (T3937?C). This transition will result into
an e3r haplotype. Finally (pathway C), a mutation occurring
in an e4 haplotype at nucleotide 4,075 changing a C to a T
(C4075?T) cannot be excluded. This transition will result into
an e3r haplotype. Notably, whereas the intragenic recombi-
SERIPA ET AL.
nation (pathway A) may be difficult to be considered in an
evolutionary mechanism, the two single-nucleotide varia-
tions (pathway B and C) may have important implications.
The Evolution of the Common APOE Haplotypes
Up to date, several hypotheses have been made regard-
ing the evolution of the common APOE variants. However,
none of these hypotheses takes into account that these
variants were four and not three.20,29–31 Here we have re-
interpreted the commonly accepted hypothesis for the
evolution of the common APOE gene variants, in which
we include the e3r haplotype, which was previously not
considered.
In the APOE gene, both SNPs rs429358 and rs7412 lie
within CpG islands, which are ‘‘hot-spot’’ mutation sites for
the methylation of cytosine and for its deamination to thy-
mine (CytosineþCH3?5mCytosine?Thymine þ NH3).18,30
This is a common process naturally occurring in DNA. In
fact, both codons 112 and 158, encoding Arg in the E4 iso-
form, are CGC codons accordingly to 2/3 of the arginine
codons in the APOE gene, and to the more frequent arginine
codons in the APOE gene being CGX.18 Assuming this pro-
cess as naturally occurring in DNA during the evolution, the
e4 haplotype (C3937-C4075) appears the more unstable,
whereas haplotypes e3 (T3937-C4075) and e3r (C3937-T4075) ap-
pear to have an intermediate stability, and e2 (T3937-T4075) the
more stable. Accordingly, the e4 haplotype plays the role of
an ancestral haplotype well; e3 and e3r play the role of
transitional haplotypes, whereas e2 well play the role of a
final product of the evolutionary course.
 Table 3. The Less Common Apolipoprotein E Isoform

Codon 112 Codon 158 Genotype Phenotype Additional change/s Trivial name Reference (see Appendix 1.)

Cys Arg e3 E3 — E3 Emi et al. (1988)19

3
Cys Arg e3 E5 Glu ?Lys E5 Maeda et al. (1989)
Cys Arg e3 E5 Glu13?Lys E5 Mailly et al. (1991)
Cys Arg e3 E4 Glu13?Lys, Arg145?Cys E4 Philadelphia Lohse et al. (1991)
Cys Arg e3 E2 Arg25?Cys E2 Kyoto Matsunaga et al. (1999)
Cys Arg e3 E2 Arg114?Cys E Tsukuba Hagiwara et al. (2008)
Cys Arg e3 E1 Gly127?Asp E1 (Weisgraber allele) Weisgraber et al. (1984)
Cys Arg e3 E2 Arg145?Pro E Sendai Oikawa et al. (1997)
Cys Arg e3 E1 Lys146?Glu E1 (variant) Moriyama et al. (1992)
Cys Arg e3 E2 Lys146?Gln E2 (variant) Rall et al. (1983)
Cys Arg e3 E1 Lys146?Asn, Arg147?Trp E1 Hammersmith Hoffer et al. (1996)
Cys Arg e3 Not reported Arg147?Pro E Chicago Sam et al. (2006)
Cys Arg e3 Not reported Arg150?Pro E Guanghzou Luo et al. (2008)
Cys Arg e3 Not reported Ala152?Asp E Las Vegas Bomback et al. (2010)
Cys Arg e3 E7 Glu244?Lys, Glu245?Lys E7 (E-Suita) Maeda et al. (1989); Tajima et al. (1989)
Cys Arg e3 E2 Arg228?Cys E2 Dunedin Wardell et al. (1990)

495
Cys Arg e3 Abnormal Arg251?Gly E3 (variant) Richard et al. (1997)
Cys Arg e3 E1 141–143 Del E Tokio Konishi et al. (1999)
Cys Arg e3 Not E2, E3, E4 481–489 Del E Maebashi Ogawa et al. (2000)
Cys – – E1 156–173 Del E1 (variant) Ando et al. (1999)
Cys Arg e3 E1 Not reported E1 Harrisburg Mann et al. (1989)
Cys Arg158?Cys e2 E2 – E2 Emi et al. (1988)19
Cys Arg158?Cys e2 E2 Arg136?Ser E2 Christchurch Wardell et al. (1987)
Cys Arg158?Cys e2 E1 Arg180?Cys E1 Baden Hoffmann et al. (2001)
Cys Arg158?Cys e2 Not reported Arg150?Gly E Okayama Kinomura et al. (2008)
Cys Arg158?Cys e2 E1 Not reported E Bethesda Gregg et al. (1983)
Cys Arg158?Cys e2 E2 Arg224?Gly E2 Fukuoka Moriyama et al. (1996)
Cys112?Arg Arg e4 E4 — E4 Emi et al. (1988)
Cys112?Arg Arg e4 Not reported Leu28?Pro E4 Pittsburgh/E4 Freiburg Kamboh et al. (1999)
Cys112?Arg Arg e4 E3 Leiden 121–127 Dupl E3 Leiden Wardell et al. (1989)
Cys112?Arg Arg158?Cys e3r E3 — E3R (E1y) Persico et al. (2004)21; Murrel et al. (2006)22;
Seripa et al. (2007)20

Cys, Cysteine; Arg, arginine; Glu, glutamic acid; Lys, lysine; Gly, glycine; Asp, aspartic acid; Pro, proline; Gln, glutamine; Asn, asparagine; Trp, tryptophan; Ala, alanine; Del, deletion; Ser, serine; Leu,
leucine; Dupl, duplication.
496
FIG. 6. Mutations possibly leading to the formation of an e3r haplotype.
Indeed, this is the hypothesis commonly accepted re-
garding the evolution of the common APOE gene variants.31
Assuming the e4 haplotype to be the ancestral form (Fig. 7),
in the course of evolution, an initial C?T transition at po-
sition 3,937 (C3937?T), favored by the methylation of a cy-
tosine, should give rise to the e3 haplotype, and a second
C?T transition at 4,075 (C4075?T), favored by the same
mechanism, will produce the e2 haplotype (pathway A).
Otherwise, assuming the same selective pressure acting at
nucleotide 3,937 as well as at nucleotide 4,075, an initial C?T
transition at position 4.075 (C4075?T), favored by the meth-
ylation of a cytosine, should give rise to the e3r haplotype,
and a second C?T transition at 3,937 (C3937?T), by the same
mechanism, will produce the e2 haplotype (pathway B).
This pathway explains the presence of e3r and its role as a
transitional haplotype. It was missing in the previous stud-
SERIPA ET AL.
ies.28–30 This favorable selective pressure from e4 toward e2
may be further supported by the unfavorable metabolic
status associated with the e4 haplotype and by a favorable
metabolic status associated with the e2 haplotype.2,32 This
hypothesis assumes an identical role for the e3 and e3r
haplotypes, also according to the identical physical charac-
teristics showed by the encoded protein isoforms. It remains
to be explained why we currently observed a major fre-
quency of e3 in respect to e3r. Indeed, this suggests that
pathways A and B in Fig. 7 are not equivalent.
The Rarity of the e3r Haplotype
The assumption that both nucleotide positions 3,937 and
4,075 have undergone the same selective pressure may be not
correct. A different selective pressure at nucleotide positions
THE FOUR APOE ALLELES
FIG. 7. Currently accepted hypothesis for the evolution of the APOE haplotypes. In its original formulation (pathway A),
this hypothesis assumes the e4 as the ancestral form, but it did not consider the existence of the e3r haplotype. The inclusion of
the e3r haplotype (pathway B) suggests an alternative evolutionary strategy that leads from the e4 to the e2 haplotype.
3,937 and 4,075 may be supported by a selective advantage
associated with the presence of cysteine at position 112
(Cys112), whereas the presence of cysteine at position 158
(Cys158) may be not advantageous.33 This was also suggested
by an in vitro experiment that indicated a possible contri-
bution of Cys158 to b-amyloid binding and by an hypercho-
lesterolemic status described in mice carrying the e3r
haplotype.34,35 Thus, the two evolutionary pathways across
the e3 and e3r may be not equivalent. It is clear that the early
use of protein-based analysis methods for the analysis and
nomenclature of the APOE isoforms may have greatly con-
tributed to the underestimation of the frequency of the e3r
haplotype. Conversely, DNA changes not favored in evolu-
tion (Fig. 6, pathways A and B) cannot be excluded regarding
their contribution as a de novo mutation in increasing the
frequency of the e3r haplotype. However, all of these
mechanisms did not explained the great discrepancy ob-
served between the frequencies of e3 and e3r in humans.
Thus, the rarity of the e3r haplotypes still remain unclear.
Possible Clinical Implications
According to the hypothesis reported in Fig. 7, we ex-
pected a disappearance of the e4 haplotype and an increase
of the e2 haplotype throughout the e3 and e3r haplotypes.
This hypothesis is also supported by the functional role of
the apoE variants. In a given population, the alleles of a
given polymorphism strictly reflect the functional role of the
497
encoded proteins. Because to a reduced breeding, an allele
associated with a pathological status will go toward negative
selection. Conversely, because to an increased breeding, an
allele associated with a healthy status and longevity will go
toward positive selection. This may be the case for the APOE
gene variants.
Even though several studies have suggested a negative
role of the e4 haplotype on survival in the elderly,36,37 con-
firmed in different ethnic groups,38 or a significant associa-
tion with mortality dependent on AD or cognitive
impairment,39–41 other studies have reported contrasting re-
sults with respect to longevity,42 or to vascular and non-
vascular-related mortality.43 In particular, a recent study
demonstrated that the e2 haplotype is significantly associ-
ated with a reduced risk of cardiovascular mortality, without
any role for the e4 haplotype.25 One hypothesis suggests that
the role of the e4 haplotype as at-risk for mortality may de-
crease with advancing age.42,44 Nevertheless, the role of the
e2 haplotype with longevity may be explained by the fol-
lowing hypotheses. First of all, the increase of the frequency
of the e2 haplotype with advancing age in healthy elderly
people may be consequential to a progressive selection due
to the disappearance of the e4 haplotype.45 Second, it may be
possible that the e2 haplotype may be a longevity-enhancing
variant, acting along pathways that are different in respect to
the e2 and e3 pathways of action. This second hypothesis is
supported by several studies indicating that the frequency
of the e3 haplotype is independent of advancing age, as
498
expected if the increase of the e2 haplotype directly corre-
sponds to the decrease of the e4 haplotype.46
Concluding Remarks
As reported, the genetics of the common APOE gene var-
iants is neither simple nor completely clear. In the last de-
cades, the fourth haplotype e3r was missing, thus obscuring
interesting questions regarding the evolution of these com-
mon APOE haplotypes. The presence of e3r well matched
with the commonly accepted hypothesis explaining the evo-
lution of the four APOE haplotypes, and also supported by
the metabolic status associated with the e2 and e4 variants.
This hypothesis, however, does not clearly explain several
points. First of all, because the e4 haplotype is the ancestral
form, and it is associated with a number of pathological
conditions and diseases, we expected its loss in the course of
evolution, and not the second highest frequency after e3 as
currently observed in humans. Second, because the e3 and e3r
haplotypes are neutral haplotypes, and they encode proteins
with the same chemical and physical characteristics, we ex-
pected the same intermediate frequency instead of the great
discrepancy currently observed between these two haplotypes.
Currently, the e3 form appears as a haplotype, providing a
selective advantage in humans, given its higher frequency in
human races, whereas the e3r form, present in only four fam-
ilies around the world, better matches with random de novo
mutation than with a further haplotype with an important role
in the evolution. Third, because the e2 haplotype appears to be
associated to healthy aging and longevity, and it appears as a
stable form that may be the final product of the evolutionary
process, we expected it to have an increasing frequency in the
population, making it the most frequent APOE variant, and not
the less frequent variant, as currently observed in humans.
Finally, it must be remembered that any selective action, to be
effective, must occur before or during the reproductive age,
and both the e2 and e4 alleles are associated with pathological
conditions that typically occur in postreproductive ages.
However, it may be possible that the selective pressures for the
progression of e4 to e3 to e2 are unrelated to these conditions,
but are likely to be related to poorly defined roles of apoE
acting before or during the reproductive age, such as modu-
lation of immunologic responses or infectious diseases or to
unknown factors now disappeared.30,47
It is clear that the role of the APOE gene variants may
have changed during the course of the evolution. Never-
theless, a better understanding of the evolution of the most-
studied polymorphism in human races may greatly improve
our knowledge of age-related diseases as well as healthy
aging and longevity.
Acknowledgments
This work was completely supported from ‘‘Ministry of
Health’’, I.R.C.C.S. Research Program, Ricerca Corrente
2009–2011, Line n. 2 ‘‘Malattie complesse’’.
References
1. Nickerson DA, Taylor SL, Fullerton SM, Weiss KM, Clark AG,
Stengard JH, Salomaa V, Boerwinkle E, Sing CF. Sequence
diversity and large-scale typing of SNPs in the human apoli-
poprotein E gene. Genome Res 2000;10:1532–1545.
SERIPA ET AL.
2. Siest G, Pillot T, Regis-Bailly A, Leininger-Muller B, Stein-
metz J, Galteau MM, Visvikis S. Apolipoprotein E: An im-
portant gene and protein to follow in laboratory medicine.
Clin Chem 1995;41:1068–1086.
3. Seripa D, Franceschi M, Matera MG, Panza F, Kehoe PG,
Gravina C, Orsitto G, Solfrizzi V, Di Minno G, Dallapiccola
B, Pilotto A. Sex differences in the association of apolipo-
protein E and angiotensin-converting enzyme gene poly-
morphisms with healthy aging and longevity: a population-
based study from Southern Italy. J Gerontol A Biol Sci Med
Sci 2006;61:918–923.
4. Seripa D, Panza F, Franceschi M, D’Onofrio G, Solfrizzi V,
Dallapiccola B, Pilotto A. Non-apolipoprotein E and apoli-
poprotein E genetics of sporadic Alzheimer’s disease. Age-
ing Res Rev 2009;8:214–236.
5. Shore B, Shore V. Heterogeneity of human plasma very low
density lipoproteins. Separation of species differing in pro-
tein components. Biochemistry 1973;12:502–507.
6. Utermann G, Jaeschke M, Mangel J. Familial hyperlipopro-
teinemia type III: deficiency of a specific apolipoprotein (apo
E-III) in the very-low-density lipoproteins. FEBS Lett
1975;56:352–355.
7. Utermann G, Canzler H, Hees M, Jaeschke M, Mühlfellner
G, Schoenborn W, Vogelberg KH. Studies on the metabolic
defect in broad-beta disease (hyperlipoproteinaemia type
III). Clin Genet 1977;12:139–154.
8. Utermann G, Hees M, Steinmetz A. Polymorphism of apo-
lipoprotein E and occurrence of dysbetalipoproteinemia in
man. Nature 1977;269:604–607.
9. Pagnan A, Havel RJ, Kane JP, Kotite L. Characterization of
human very low density lipoproteins containing two elec-
trophoretic populations: Souble pre-beta lipoproteinemia and
primary dysbetalipoproteinemia. J Lipid Res 1977;18:613–
622.
10. Marcel YL, Bergseth M, Nestruck AC. Preparative isoelectric
focusing of apolipoproteins C and E from human very low
density lipoproteins. Biochim Biophys Acta 1979;573:175–
183.
11. Warnick GR, Mayfield C, Albers JJ, Hazzard WR. Gel fo-
cusing method for specific diagnosis of familial hyperlipo-
proteinemia type 3. Clin Chem 1979;25:279–284.
12. Weidman SW, Suarez B, Falko JM, Witztum JL, Kolar J,
Raben M, Schonfeld G. Type III hyperlipoproteinemia: De-
velopment of a VLDL ApoE gel isoelectric focusing tech-
nique and application in family studies. J Lab Clin Med
1979;93:549–569.
13. Zannis VI, Breslow JL. Characterization of a unique human
apo E variant associated with type III hyperlipoproteinemia.
J Biol Chem 1980;255:1759–1762.
14. Zannis VI, Just PW, Breslow JL. Human apolipoprotein E
isoprotein subclasses are genetically determined. Am J Hum
Genet 1981;33:11–24.
15. Zannis VI, Breslow JL, Utermann G, Mahley RW, Weis-
graber KH, Havel RJ, Goldstein JL, Brown MS, Schonfeld G,
Hazzard WR, Blum C. Proposed nomenclature of apoE
isoproteins, apoE genotypes, and phenotypes. J Lipid Res
1982;23:911–914.
16. Breslow JL, McPherson J, Nussbaum AL, Williams HW,
Lofquist-Kahl F, Karathanasis SK, Zannis VI. Identification
and DNA sequence of a human apolipoprotein E cDNA
clone. J Biol Chem 1982;257;14639–41461.
17. Das HK, McPherson J, Bruns GA, Karathanasis SK, Breslow
JL. Isolation, characterization, and mapping to chromosome
THE FOUR APOE ALLELES
19 of the human apolipoprotein E gene. J Biol Chem
1985;260:6240–6247.
18. Paik YK, Chang DJ, Reardon CA, Davies GE, Mahley RW,
Taylor JM. Nucleotide sequence and structure of the human
apolipoprotein E gene. Proc Natl Acad Sci USA 1985;82:3445–
3449.
19. Emi M, Wu LL, Robertson MA, Myers RL, Hegele RA, Williams
RR, White R, Lalouel JM. Genotyping and sequence analysis of
apolipoprotein E isoforms. Genomics 1988;3:373–379.
20. Seripa D, Matera MG, Daniele A, Bizzarro A, Rinaldi M,
Gravina C, Bisceglia L, Corbo RM, Panza F, Solfrizzi V,
Fazio VM, Forno GD, Masullo C, Dallapiccola B, Pilotto A.
The missing ApoE allele. Ann Hum Genet 2007;71:496–500.
21. Persico AM, D’Agruma L, Zelante L, Militerni R, Bravaccio C,
Schneider C, Melmed R, Trillo S, Montecchi F, Elia M, Palermo
M, Rabinowitz D, Pascucci T, Puglisi-Allegra S, Reichelt KL,
Muscarella L, Guarnieri V, Melgari JM, Conciatori M, Keller F.
Enhanced APOE2 transmission rates in families with autistic
probands. Psychiatr Genet 2004;14:73–82.
22. Murrell JR, Price BM, Baiyewu O, Gureje O, Deeg M, Hen-
drie H, Ogunniyi A, Hall K. The fourth apolipoprotein E
haplotype found in the Yoruba of Ibadan. Am J Med Genet B
Neuropsychiatr Genet 2006;141B:426–427.
23. http://browser.1000genomes.org
24. Hendrie HC, Ogunniyi A, Hall KS, Baiyewu O, Unverzagt
FW, Gureje O, Gao S, Evans RM, Ogunseyinde AO,
Adeyinka AO, Musick B, Hui SL. Incidence of dementia and
Alzheimer disease in 2 communities: Yoruba residing in
Ibadan, Nigeria, and African Americans residing in In-
dianapolis, Indiana. JAMA 2001;285:739–747.
25. al-Chalabi A, Enayat ZE, Bakker MC, Sham PC, Ball DM,
Shaw CE, Lloyd CM, Powell JF, Leigh PN. Association of
apolipoprotein E epsilon 4 allele with bulbar-onset motor
neuron disease. Lancet 1996;347:159–160.
26. Matera MG, Sancarlo D, Panza F, Gravina C, D’Onofrio G,
Frisardi V, Longo G, D’Ambrosio LP, Addante F, Copetti M,
Solfrizzi V, Seripa D, Pilotto A. Apolipoprotein E-related all-
cause mortality in hospitalized elderly patients. Age (Dordr)
2010;32:411–420.
27. Seripa D, Signori E, Gravina C, Matera MG, Rinaldi M, Fazio
VM. Simple and effective determination of apolipoprotein E
genotypes by positive/negative polymerase chain reaction
products. Diagn Mol Path 2006;15:180–185.
28. Iurescia S, Fioretti D, Mangialasche F, Rinaldi M. The path-
ological cross talk between apolipoprotein E and amyloid-
beta peptide in Alzheimer’s disease: Emerging gene-based
therapeutic approaches. J Alzheimers Dis 2010;21:35–48.
29. Hanlon CS, Rubinsztein DC. Arginine residues at codons
112 and 158 in the apolipoprotein E gene correspond to the
ancestral state in humans. Atherosclerosis 1995;112:85–90.
30. Finch CE, Sapolsky RM. The evolution of Alzheimer disease,
the reproductive schedule, and apoE isoforms. Neurobiol
Aging 1999;20:407–428.
31. Mahley RW, Rall SC Jr. Is epsilon 4 the ancestral human
apoE allele? Neurobiol Aging 1999;20:429–430.
32. Panza F, D’Introno A, Colacicco AM, Capurso C, Capurso S,
Kehoe PG, Capurso A, Solfrizzi V. Vascular genetic factors
and human longevity. Mech Ageing Dev 2004;125, 169–178.
33. Corbo RM, Scacchi R. Apolipoprotein E (APOE) allele dis-
tribution in the world. Is APOE*4 a ’thrifty’ allele? Ann Hum
Genet 1999;63:301–310.
34. Aleshkov SB, Li X, Lavrentiadou SN, Zannis VI. Contribu-
tion of cysteine 158, the glycosylation site threonine 194, the
499
amino-and carboxy-terminal domains of apolipoprotein E in
the binding to amyloid peptide beta (1–40). Biochemistry
1999;38:8918–8925.
35. Role of apoE in Lipoprotein Metabolism and Type III HLP.
Accessed at http://www.gladstone.ucsf.edu/gladstone/
site/Mahley/content/1/70, February 2, 2011.
36. Louhija J, Viitanen M, Aguero-Torres H, Tilvis R. Survival in
Finnish centenarians in relation to apolipoprotein E poly-
morphism. J Am Geriatr Soc 2001;49:1007–1008.
37. Tilvis RS, Strandberg TE, Juva K. Apoliprotein E pheno-
types, dementia and mortality in a prospective population
sample. J Am Geriatr Soc 1998;46:712–715.
38. Ewbank DC. Differences in the association between apoli-
poprotein E genotype and mortality across populations. J
Gerontol A Biol Sci Med Sci 2007;62:899–907.
39. Dal Forno G, Carson KA, Brookmeyer R, Troncoso J, Kawas
CH, Brandt J. APOE genotype and survival in men and wo-
men with Alzheimer’s disease. Neurology 2002;58:1045–1050.
40. Rosvall L, Rizzuto D, Wang HX, Winblad B, Graff C, Fratiglioni
L. APOE-related mortality: Effect of dementia, cardiovascular
disease and gender. Neurobiol Aging 2009;30:1545–1551.
41. Rockwood K, Nassar B, Mitnitski A. Apolipoprotein E-
polymorphism, frailty and mortality in older adults. J Cell
Mol Med 2008;12:2754–2761.
42. Gerdes LA, Klausen IB, Sihm I, Færgeman O. Apolipopro-
tein E polymorphism in a Danish population compared to
findings in 45 other study populations around the world.
Genet Epidemiol 1992;9:155–167.
43. Panza F, D’Introno A, Capurso C, Colacicco AM, Seripa D,
Pilotto A, Santamato A, Capurso A, Solfrizzi V. Lipopro-
teins, vascular-related genetic factors, and human longevity.
Rejuvenation Res 2007;10:441–458.
44. Olichney JM, Sabbagh MN, Hofstetter CR, Galasko D,
Grundman M, Katzman R, Thal LJ. The impact of apolipo-
protein E4 on cause of death in Alzheimer’s disease. Neu-
rology 1997;49:76–81.
45. Ewbank DC. Mortality differences by APOE genotype esti-
mated from demographic synthesis. Genet Epidemiol 2002;
22:146–155.
46. Koivisto AM, Lempiäinen P, Koivisto K, Helkala EL, Myk-
känen L, Kuusisto J, Kervinen K, Kesäniemi YA, Laakso M,
Soininen H. Apolipoprotein E phenotype alone does not in-
fluence survival in Alzheimer’s disease: A population-based
longitudinal study. Neuroepidemiology 2000;19:327–332.
47. Mahley RW. Apolipoprotein E: Cholesterol transport protein
with expanding role in cell biology. Science 1988;240:622–
630.
Address correspondence to:
Davide Seripa
Gerontology & Geriatrics Research Laboratory
Department of Medical Sciences
I.R.C.C.S. ‘‘Casa Sollievo della Sofferenza’’
San Pio da Pietrelcina Foundation
Viale Cappuccini 1
71013 San Giovanni Rotondo (FG)
Italy
E-mail: dseripa@operapadrepio.it
Received: February 11, 2011
Accepted: April 11, 2011
Appendix follows?
500 SERIPA ET AL.

Appendix 1.

References for Table 3.

Ando M, Sasaki J, Hua H, Matsunaga A, Uchida K, Jou K, Oikawa S, Saito T, Nihei H. A novel 18-amino acid deletion in
apolipoprotein E associated with lipoprotein glomerulopathy. Kidney Int. 1999;56:1317–1323.
Bomback AS, Song H, D’Agati VD, Cohen SD, Neal A, Appel GB, Rovin BH. A new apolipoprotein E mutation, apoE Las
Vegas, in a European-American with lipoprotein glomerulopathy. Nephrol Dial Transplant. 2010;25:3442–3446.
Gregg RE, Ghiselli G, Brewer HB Jr. Apolipoprotein EBethesda: a new variant of apolipoprotein E associated with type III
hyperlipoproteinemia. J Clin Endocrinol Metab. 1983;57:969–974.
Hagiwara M, Yamagata K, Matsunaga T, Arakawa Y, Usui J, Shimizu Y, Aita K, Nagata M, Koyama A, Zhang B, Mastunaga
A, Saku K, Saito T. A novel apolipoprotein E mutation, ApoE Tsukuba (Arg 114 Cys), in lipoprotein glomerulopathy.
Nephrol Dial Transplant. 2008;23:381–384.
Hoffer MJ, Niththyananthan S, Naoumova RP, Kibirige MS, Frants RR, Havekes LM, Thompson GR. Apolipoprotein
E1-Hammersmith (Lys146–>Asn;Arg147–>Trp), due to a dinucleotide substitution, is associated with early manifestation
of dominant type III hyperlipoproteinaemia. Atherosclerosis. 1996;124:183–189.
Hoffmann MM, Scharnagl H, Köster W, Winkler K, Wieland H, März W. Apolipoprotein E1 Baden (Arg(180)–>Cys). A new
apolipoprotein E variant associated with hypertriglyceridemia. Clin Chim Acta. 2001;303:41–48.
Kamboh MI, Aston CE, Perez-Tur J, Kokmen E, Ferrell RE, Hardy J, DeKosky ST. A novel mutation in the apolipoprotein E gene
(APOE*4 Pittsburgh) is associated with the risk of late-onset Alzheimer’s disease. Neurosci Lett. 1999;263(2–3):129–132.
Kinomura M, Sugiyama H, Saito T, Matsunaga A, Sada KE, Kanzaki M, Takazawa Y, Maeshima Y, Yanai H, Makino H. A
novel variant apolipoprotein E Okayama in a patient with lipoprotein glomerulopathy. Nephrol Dial Transplant.
2008;23:751–756.
Konishi K, Saruta T, Kuramochi S, Oikawa S, Saito T, Han H, Matsunaga A, Sasaki J. Association of a novel 3-amino acid
deletion mutation of apolipoprotein E (Apo E Tokyo) with lipoprotein glomerulopathy. Nephron. 1999;83:214–218.
Lohse P, Mann WA, Stein EA, Brewer HB Jr. Apolipoprotein E-4Philadelphia (Glu13—-Lys,Arg145—-Cys). Homozygosity
for two rare point mutations in the apolipoprotein E gene combined with severe type III hyperlipoproteinemia. J Biol
Chem. 1991;266:10479–10484.
Luo B, Huang F, Liu Q, Li X, Chen W, Zhou SF, Yu X. Identification of apolipoprotein E Guangzhou (arginine 150 proline),
a new variant associated with lipoprotein glomerulopathy. Am J Nephrol. 2008;28:347–353.
Maeda H, Nakamura H, Kobori S, Okada M, Niki H, Ogura T, Hiraga S. Molecular cloning of a human apolipoprotein
E variant: E5 (Glu3—-Lys3). J Biochem. 1989; 105:491–493.
Mailly F, Xu CF, Xhignesse M, Lussier-Cacan S, Talmud PJ, Davignon J, Humphries SE, Nestruck AC. Characterization of
a new apolipoprotein E5 variant detected in two French-Canadian subjects. J Lipid Res. 1991;32:613–620.
Mann WA, Gregg RE, Sprecher DL, Brewer HB Jr. Apolipoprotein E-1Harrisburg: a new variant of apolipoprotein
E dominantly associated with type III hyperlipoproteinemia. Biochim Biophys Acta. 1989;1005:239–244.
Matsunaga A, Sasaki J, Komatsu T, Kanatsu K, Tsuji E, Moriyama K, Koga T, Arakawa K, Oikawa S, Saito T, Kita T, Doi T. A
novel apolipoprotein E mutation, E2 (Arg25Cys), in lipoprotein glomerulopathy. Kidney Int. 1999;56:421–427.
Moriyama K, Sasaki J, Matsunaga A, Arakawa F, Takada Y, Araki K, Kaneko S, Arakawa K. Apolipoprotein E1 Lys-146—-
Glu with type III hyperlipoproteinemia. Biochim Biophys Acta. 1992;1128:58–64.
Moriyama K, Sasaki J, Takada Y, Arakawa F, Matsunaga A, Ito Y, Arakawa K. Characterization of a novel variant of
apolipoprotein E, E2 Fukuoka (Arg-224 –>Gln) in a hyperlipidemic patient with xanthomatosis. Biochim Biophys Acta.
1996;1301:185–190.
Ogawa T, Maruyama K, Hattori H, Arai H, Kondoh I, Egashira T, Watanabe T, Kobayashi Y, Morikawa A. A new variant of
apolipoprotein E (apo E Maebashi) in lipoprotein glomerulopathy. Pediatr Nephrol. 2000;14:149–151.
Oikawa S, Matsunaga A, Saito T, Sato H, Seki T, Hoshi K, Hayasaka K, Kotake H, Midorikawa H, Sekikawa A, Hara S, Abe
K, Toyota T, Jingami H, Nakamura H, Sasaki J. Apolipoprotein E Sendai (arginine 145–>proline): a new variant associated
with lipoprotein glomerulopathy. J Am Soc Nephrol. 1997;8:820–823.
Rall SC Jr, Newhouse YM, Clarke HR, Weisgraber KH, McCarthy BJ, Mahley RW, Bersot TP. Type III hyperlipoproteinemia
associated with apolipoprotein E phenotype E3/3. Structure and genetics of an apolipoprotein E3 variant. J Clin Invest.
1989;83:1095-1101.
Richard P, Beucler I, Pascual De Zulueta M, Biteau N, De Gennes JL, Iron A. Compound heterozygote for both rare
apolipoprotein E1 (Gly127–>Asp, Arg158–>Cys) and E3(Cys112–>Arg, Arg251–>Gly) alleles in a multigeneration
pedigree with hyperlipoproteinaemia. Clin Sci (Lond). 1997;93:89–95.
Sam R, Wu H, Yue L, Mazzone T, Schwartz MM, Arruda JA, Dunea G, Singh AK. Lipoprotein glomerulopathy: a new
apolipoprotein E mutation with enhanced glomerular binding. Am J Kidney Dis. 2006;47:539–548.
Tajima S, Yamamura T, Menju M, Yamamoto A. Analysis of apolipoprotein E7 (apolipoprotein E-Suita) gene from a patient
with hyperlipoproteinemia. J Biochem. 1989;105:249–253.
Wardell MR, Brennan SO, Janus ED, Fraser R, Carrell RW. Apolipoprotein E2-Christchurch (136 Arg—-Ser). New variant of
human apolipoprotein E in a patient with type III hyperlipoproteinemia. J Clin Invest. 1987;80:483–490.
Wardell MR, Weisgraber KH, Havekes LM, Rall SC Jr. Apolipoprotein E3-Leiden contains a seven-amino acid insertion that
is a tandem repeat of residues 121-127. J Biol Chem. 1989;264:21205–21210.
Wardell MR, Rall SC Jr, Brennan SO, Nye ER, George PM, Janus ED, Weisgraber KH. Apolipoprotein E2-Dunedin (228 Arg
replaced by Cys): an apolipoprotein E2 variant with normal receptor-binding activity. J Lipid Res. 1990;31:535–543.
Weisgraber KH, Rall SC Jr, Innerarity TL, Mahley RW, Kuusi T, Ehnholm C. A novel electrophoretic variant of human
apolipoprotein E. Identification and characterization of apolipoprotein E1. J Clin Invest. 1984;73:1024-1033.