HLA allele and haplotype frequencies in Algerians

ELSEVIER

HLA Allele and Haplotype Frequencies in Algerians Relatedness to Spaniards and Basques

Antonio Arnaiz-Villena, Djamal Benmamar, Miguel Alvarez, Nieves Diaz-Campos, Pilar Varela, Eduardo Gomez-Casado, and Jorge Martinez-Las0

ABSTRACT: The powerful genetic polymorphism of the HLA system has been used to identify individuals and populations. Ethnic groups may be characterized by specific HLA allele frequencies and particular extended HLA haplotypes; also, genetic relationships among these groups may be deduced. In the present study, serology and DNA typing were used to detect HLA-A, -B, -C, -DR, and -DQ alleles in each individual and to calculate characteristic haplotypes in Algerians. These results were compared to those previously obtained in other populations, particularly northern Mediterraneans; genetic dis-

tances and their respective dendrograms place Basques and Spaniards closer to Algerians than to other Europeans. Also, characteristic Basque and/or Spanish haplotypes are found in Algerians; i.e., A3O-Bl8-CwT-DR3-DQ2 and A 1 -B57-Cw7-DR7-DQ2. This supports the evidence that the Algerian population, mainly its paleo-North African component (Berbers), has a common descent with Basques and Spaniards, probably reflecting a preneolithic relationship between Iberians and paleo-North Africans. Human Immunology 43) 259-268 (1995)

ABBREVIATIONS

:; haplotype frequency neighbor-joining

RLD SGD

relative linkage disequilibria standard genetic distance

INTRODUCTION

The population of Africa in pteneolithic times (about

7000-3000 years BC) has been classified into live main ethnic groups: (1) Hamites (Hamite-speaking people) at

the Mediterranean and Red Sea coasts, they were white

people who have been further subdivided into Berbers

and Egyptians (approximately n = 100,000) and whitc- black Ethiopians (n = 100,000); (2) blacks at the

Guinea Golf (n = 250,000); (3) black Nilo-Saharians (n

= 250,000), surrounding the present-day Sahara Desert and Nilo River; (4) black Pygmies at southwestern Af- rican coasts (n = 200,000); and (5) black Bushmen at

south and southeast coasts (n = 350,000) (see Fig. 1A)

El}. In Pharaonic Africa (3000 years BC) population

numbers had suffered drastic changes. Agricultural Egypt had 1 million people and the rest of Africa (non-

agricultural hunters) had 1,200,OOO altogether. By 400

years AD, blacks from Guinea Golf had overwhelmed

central and southern Africa, reduced Bushmen, and al-

most eliminated Pygmies (see Fig. 1B) El].

From the Department of Immunology, University Complutense, Hospital 12 de Octubre, Cawetera Anablucia, Madrid, Spain.

Address reprint requests to A. Arnaiz-Villena, Departamento de Inmu- nologia, Universihd Comphtense, Hospital I2 de Octthe, Cawetera An- Llucia, 28041 Madrid, Spain.

Drastic climatic changes had dried northern Africa by

around 8000-4000 years BC; the forest line had gone down towards the Equator from about parallel 16 by

3000 years BC (Fig. 1C) 121. By then the Sahara Desert had similar characteristics to the present-day Sahara and

strong migrations towards the north sent people to the Iberian Peninsula and Mediterranean and Canary Islands; the first Iberians and Canary Islands “guanches” may in part come from these displaced paleo-North African in-

habitants (Hamite-Berbers) or be closely related (33.

Receivedjuly 19, 1994; accepted February 6, 1995.

The population of Algeria has followed the general Maghteb demographic movements El). In neolithic

Human Immunology 43, 259-268 (1995) 0198-8859l95lS9.50 0 American Society for Histocompatibility and Immunogenetics, 1995 SSDI 0198-8859(95)00024-X

260

FIGURE 1 African Population changes between 3000 years BC and 400 years AD. (A) Ethnic groups in Africa 3000 years BC. (B) Ethnic groups in Africa 400 years AD. (C) Line of jungle displacement between 8000 and 1000 years BC.

times a scattering of Berber pastoralists and cultivators (a

few hundred thousand) existed there. However, they re- mained stuck at a simple neolithic level while other Mediterraneans were evolving through the Bronze and Iron Ages. By 1000 years BC, Phoenicians from Lebanon found a stone age neolithic culture in Maghreb. They set up Carthage (Tunisia) and were overthrown by the Ro- mans in 146 BC. At this time there were 100,000 Phoe-

nicians and 500,000 Berbers in Tunisia plus another 2.5 million Berbers in the rest of North Africa. By 200 AD

Algeria had no more than 2 million people, which were slowly going into agriculture due to Roman rule. During the seventh century Arabs invaded Algeria and mixed

A. Arnaiz-Villena et al.

with Berbers at a rate of about 40% Arab and 60%

Berber [4}. Culturally and demographically Algeria (and

Maghreb) has been stuck until very recently, when France and Spain left their colonies. The Algerian pop-

ulation was 9 million in 1950 and is more than 25

million at present.

One of the problems of Mediterranean ethnography has been the lack of genetic marker data coming from

southern Mediterranean populations [5]. The powerful and discriminating HLA polymorphism [3} has not been

studied in the Algerian population even in the last In-

ternational Workshop anthropological section 161, where

a few French Algerians were studied only by serology and

complete data were not shown. This has led to a lack of comparative genetic studies between northern and south-

ern Mediterraneans.

In the present work, we tried to fulfill this gap by

collecting blood samples from unrelated Algerians at Al- giers and testing them for HLA-A, -B, -DR, and -DQ

antigens both by serology and DNA sequencing. We also compared our data with that of Europeans, partic-

ularly Spaniards, Basques, and Sardinians, obtaining the corresponding genetic distances and a relatedness den-

drogram .

MATERIALS AND METHODS

Population Sample

One hundred six unrelated healthy Algerian blood do- nors were randomly selected as a representative panel of

this population. Blood was collected in Algiers, Hopital Central de I’ArmCe, Blood Bank. Peripheral blood lym-

phocytes were isolated by Ficoll gradient, frozen down,

and kept in liquid nitrogen until use in Madrid. Subjects did not have recorded admixture with Europeans over at

least three generations.

HLA Class I and II Serotyping

HLA class I (A, B, C) and class II (DR, DQ) typing was done by a two-step microlymphocytotoxicity technique

on T or B lymphocytes E71, respectively, using the 10th and 1 lth International Histocompatibility Workshop

and local reagents.

HLA-DRBl, DRB3, DRB4, DQAl, and DQBl Exon-2 Sequences

DNA extraction. Genomic DNA from whole blood containing EDTA was extracted with a 340A Nucleic Acid

Extractor (Applied Biosystems, Foster City, CA, USA) with the use of ABI reagents and protocols 181.

Amplification of genomic DNA. A generic HLA-DRB amplification as well as group-specific amplifications of the DR 1, DR2, DR4, DRB 1 alleles associated with DR52 and DRB3 were made. The DRB 1, DRB3, DQAl, and

HLA in Algerians 261

DQBl exon-2 DNA was amplified by polymerase chain

reaction (PCR) using Taq polymerase (Cetus, USA), as

previously described 19). Primers and oligonucleotides for DRBl, DRB3, DRB4, DQAl, and DQBl were

those described by Kimura and Sasatuki El01 for the 1 lth International Histocompatibility Workshop. They

were synthesized in a DNA-SM automated synthesizer (Beckman, Palo Alto, CA, USA).

Dot-blot hybridization. Five percent of the amplified

DNA was denatured in 0.4 mol/l NaOH for 10 minutes,

neutralized in 1 mol/l ammonium acetate, and trans-

ferred to a Hybond-N membrane (Amersham, Bucks,

UK). The filters were prehybridized for 1 hour in a solution containing 50 mmol/l Tris-HCI (pH = 8.0), 3

mol/l tetramethylammonium chloride, 2 mmolll EDTA

(pH = 8.0), 5X Denhart’s solution (2% bovine serum

albumin, 2% polyvinyl pyrrolidone 40, 2% ficoll 4OO),

and 0.1 mg/ml salmon sperm DNA. Then the radiola- beled oligonucleotide was added and hybridized at 54°C

for 3 hours. The filters were washed twice in 2X SSPE

(30X SSPE: 4.5 mol/l NaCl, 0.3 mol/l NaH,P04, 30 mmol/l EDTA, pH = 7.7) and 0.1% SDS at room

temperature for 10 minutes, once in TMAC solution (50

mmol/l Tris-HCl, pH = 8.0, 3 mol/l tetramethylammonium chloride, and 2 mmol/l EDTA, 0.1% SDS) at

room temperature for 10 minutes, and twice in TMAC

solution at 58°C for 10 minutes. The dots were visual-

ized after 8 hours of exposure to Kodak XAR5 film at - 80°C with intensifying screens.

Oligonucleotide probes. Information about the sequences

and specificities of the DRB 1, DRB3, DQA 1, and DQB 1 oligonucleotides are from the 1 lth International

Histocompatibility Workshop [ 101. Oligonucleotide

synthesis was done using the cyanoethyl phosphora-

midite technique in a Beckman DNA-SM automated

DNA synthesizer, following the manufacturer’s protocol. The oligonucleotides were labeled and purified ac-

cording to the methods of Nepom et al. Ill].

Cloning of amplified DNA and automated dideoxy sequencing. Amplification products were purified and inserted

into the Replicative Form of SmaI-digested M13mp18 virus DNA (Boehringer Mannheim, Germany). Trans- formation of competent Eschericbia coli TG 1 cells, screen-

ing for HLA-DRB inserts, and purification of the recom-

binants was done as previously described [9]. Single- stranded DNA templates were sequenced using Sanger’s dideoxy chain termination method on an Applied Bio- systems 370A automated DNA sequencer {12) as de-

scribed [91. DNA from individuals with newly found DR-DQ associations was cloned and sequenced (three clones at least) in order to confirm associations.

Statistic Analysis

The linkage disequilibria between two alleles and the

level of significance (p) for 2 X 2 comparisons were

computed using the formulas of Mattiuz et al. { 131. The linkage disequilibria for comparison of more than two

alleles were calculated using the 1 lth International His-

tocompatibility Workshop methodology E 141. Also, the relative linkage disequilibria (RLD) were calculated as described in Imanishi et al. f 141 and Degos (151. Stan-

dard errors are worked out as detailed in Pickbourne et

al. { 161. The HLA-A, -B, -C, -DRBl, -DRB3, -DQAl, and -DQB 1 allelic frequencies and linkage disequilibria

were calculated by using two Fortram IV computer programs 117). In addition (see Table 6), most frequent

complete haplotypes were tentatively deduced from (a)

the 2, 3, and 4 HLA loci haplotype frequencies depicted in Tables 3, 4, and 5; (b) the previously described hap-

lotypes in other populations 118, 191; and (c) it was

helpful to define haplotypes, if they appear in 2 or more individuals and the alternative haplotype was well de-

fined. This logical combination is more accurate for the

class II loci DRB 1, DRB3, DQA 1, and DQB 1, but is

somewhat less dependable for HLA-A-B, DR or HLA-

A,B. For comparison of phenotype and haplotype frequencies the reference tables of the 11th International

Histocompatibility Workshop were used { 18).

Dendrograrns

Phylogenetic trees (dendrograms) were constructed with

the genotype frequencies by using the neighbor-joining (NJ) method E201 with the standard genetic distances

(SGDs) between populations 1211 using the software DISPAN containing the programs GNKDST and

TREEVIEW.

RESULTS AND DISCUSSION

Characteristic HLA Alleles That Identify the Algerian Population (see Tables 1 and 2)

The HLA system has been used to identify individuals in

order to solve paternity disputes and also in forensic problems [22, 231. The discriminative strength of the

HLA system stems from the fact that the number of alleles for each loci is very high. HLA strength for iden-

tifying individuals for paternity testing is higher than the additive strength of many others enzymes and blood group polymorphic systems. Presently, DNA-typed

HLA class II alleles and the discovery of new loci even increase the HLA strength for distinguishing individuals (241. This HLA discriminating ability may also be used to identify populations; particular alleles only appear or

262 A. Amaiz-Villena et al.

TABLE 1 Phenotype frequencies of HLA-A, -B, and DRB1*0102. (b) From Spaniudr: decreased: A29, B44, -C alleles in an Algerian population and DRB1’07; increased: B49 and B38 137. In sum-

Phenotype Phenotype mary, significant decreases in A29, B44, and

frequencies frequencies DRB lx0 10 1 frequencies, which are characteristic Cau-

Alleles n = 102 (%) Alleles n = 102 (%) casoid alleles, were observed 1187. Also, high frequencies

of B49, B35, and B38 were present and similar to what 23 9 43

has been found in Mediterraneans [18); B53 was also 5

16 9 increased and found also in high frequency in Negroids

12 1181.

Al A2 A3 A23 A24 A25 A26 All A28 A29 A30 A31 A32 A33 A34 A66

B51 B52 B53 B7 B8 B44 B45 B13 B14 B62 B63 B57 B58 B18 B49 B50

855 B27

Cwl 2 cw2 9 cw3 6 cw4 21 cw5 11 Cw6 17 cw7 31 Cw8 4

18 1 5

12 6 5

II 6

10 9 3 1

17 4

11 5 2

10 5

19 10 2 2

B35 20 B37 2 B38 12 B39 3 B60 2 B61 4 B41 3 B59 1 B70 3 B73 1

are very frequent in certain populations (i.e., A36, A43

in Negroids) and the strong linkage disequilibrium between HLA neighboring loci shows that certain combi-

nations of neighboring alleles (HLA haplotypes) are in characteristic frequency (or absolutely) in certain living

populations. Once the HLA ability to identify populations has been

estabiished, it is a unique tool for studying the origins of

relatively isolated groups, as the characteristic HLA allele frequencies have not been completely diluted out with time. Mixing HLA allele frequencies with another

system’s allele frequencies coming from nondiscrimina-

tive low polymorphic genetic systems (i.e., ABO, MN, Rh) would tend to homogenize populations and make the HLA analysis less discriminative.

In the present work, it was found that the following HLA-A, -B, and -DR frequencies of the Algerian population do differ significantly @ C 0.05; see Tables 1 and 2 and {37): (a) From Baqza: decreased: A29, B44, and DRBl*OlOl; increased: A24, B53, B49, B35, B38, and

TABLE 2 Phenotype frequencies of HLA-DRB 1, -DRB3, -DQAl and DQB 1 alleles in an Algerian population

Alleles

Phenotype frequencies

n = 106 (%) Alleles

Phenotype frequencies

n = 106 (%)

DRB 1 0101 0102 0103 1501 1502 1601 1602 0301 0302 0401 0402 0403 0404 0405 0406 0407 0408 1101 1102 1103 1104 1201 1202 1301 1302 1303 1401

07 0801 0803 0804 0806

09 1001

3 16 0

22 6 6 0

23 3 0 8 8 4 5 4 0

0

10 6 0 6

8 5 4

24 2

2 3 0

DRB3 0101 0201 0202 0301

DQA 1 01

0102 0103 0201

03 040 1 050 1 060 1

DQB 1 0501 0502 0503 0504 060 1 0602 0603 0604 0605 0201 0301 0302 0303 0402

19

40 8

38 31 11 23 28

47 1

26 8 4 0

24

7

42 29 23

0 9

Alleles DQA1*03011, 03012, and 0302 were all assigned as DQA1+03; DQB1.05031 and 05032 as DQB1+0503 and DQBP03031; and 03032 as DQB190303. Alleles DQAl+OlOl and 0104 were ail assigned as DQAl*Ol. 1 Ich International Histocompatibility Workshop nomemlacure is used.


TABLE 3 HLA-A/B, B/DRB 1, DRB l/DQB 1, and DQAl/DQB 1 two-loci haplotypes with significant linkage disequilibrium in an Algerian population

HF (%) LD x 100 RLD Haplotype HF (%) LD X 100 RLD

HLA-A-B Al-B57 A26-B44 A30-B 18 A33-B 14

B-DRB 1

B7-1001 B8-07

B 14-O 102 B18-0301 B41-1303 B44-0402 B49- 1102

B57-07 B60- 1202 Bbl-1601

DQA 1-DQB 1

01-0501

01-0503 0102-0502 0 102-0602

0103-0601 0103-0603 0201-0201

03-0302 03-0402

040 l-0402 0501-0201

0501-0301

2.9” 1.5b 1.5b

2.5”

1.9b

2.5” 3.4”

2.5” o.9b 1.9b 1.9b 3.4”

o.5b o.9b

13.2”

1.9b 4.2”

10.8”

3.3” 3.3”

9.3” 10.8”

3.3” 2.4”

12.7“ 13.7”

2.2 0.49’ 1.2 5.22’ 1.2 0.43’

2.2 0.51’

1.7 0.53d

1.9 0.73’

2.9 0.57f

1.9 0.73‘

0.9 0.64 1.4 0.3Y 1.6 0.59’ 2.7 0.61d

0.5 1.02’

0.9 0.47’

11.2 l.OY

1.5 1.00‘

3.6 l.O$f

9.3 0.88

3.2 0.97’

3.2 0.97’

7.3 0.78”

9.6 0.9Y 2.4 0.62‘

2.3 l.OV

5.1 0.29’ 10.2 0.89’

DRB l-DQB 1

0101-0501 0 102-050 1 1501-0602 1502-0601

160 l-0502

030 l-020 1 0302-0402 0402-0302

0403-0302 0405-0302 0406-0402 1101-0301 1102-0301

1104-030 1

130 l-0603

1302-0604 1303-0301

1401-0503 07-020 1

1001-0501

l.4b 1.2 0.92’ 8.0” 7.2 1.01’

9.9” 8.9 0.89’ 2.8” 2.8 l.O(Y

2.8” 2.7 0.98 11.3” 9.2 0.81’

l.4b 1.4 l.Od 4.2” 3.8 0.90h 2.8” 2.4 0.86 2.4” 2.1 0.88’ 1.9b 1.8 0.95h 4.7” 3.9 0.8g 2.8” 2.4 0.86d 2.8” 2.4 0.86d

3.3” 3.2 0.971

3.3” 3.2 0.971 2.4” 2.0 0.83d 1.9b 1.9 l.O(Y

9.9” 7.3 0.74x

3.3” 2.9 0.8s’

HF, haplotype frequency expressed ar percentage; LD, linkage disequilibrium obtained by Mattiuz formulas (131; and RLD, relative linkage disequilibrium [IdI.

Alleles DQA 1+0 10 1 and 0 104 were all assigned as DQA l*O 1. Comparisons between HLA-A/B and B-DRB l* were made in 102 Algerians DRB 1 l -DQB 1’ and DQAl*-DQBI’ were made in 106 Algerians.

Standard deviation (SD) values ate expressed as ’ = 0.010 G SD G 0.025 and 6 0.004 < SD < 0.010. ‘p < 0.05; “p < 0.005; ‘p < 0.0005;‘p < 0.00005; "p < 0.000005; "p < 0.0000005; 'p < 0.000000005; ‘p < 0.0000000005.

The < and "p V&RS do not reach significance when they ate corrected by the number of haplotypes tested 0’ = 204 or 212).

Characteristic HLA Linkage Disequilibria Found in Algerians (see Tables 3, 4, and 5)

HLA-AIB associations.

Al-B57-Cw7 (DR7-DQ2). A 1-B57-Cw6 haplotype is found in many populations over the world; however, (Al)-B57-Cw7 is only found in Spanish-Basques (hap-

lotype frequency [HF]: 1.2) 131, French-Basques (HF: 2.5), Sardinians (HF: 2.3), Hottentotes (HF: 5.4), and

Algerians (HF: 1.9) 118). This characteristic also groups

Basques with Algerians, Sardinians, and other African populations (see dendrogram in Fig. 2, p 629 in 1251).

Moreover, this Mediterranean haplotype goes with DR7 and DQ2, whereas B57-Cw6 goes with DR7-DQ9. In addition, Al-B57-Cwb-DR7 has also been found in Al-

gerians, but associated to DQS and not to DQ9, break- ing the normal Caucasoid association (Table 4). This

DR7-DQ8 association has been found in other Middle East Mediterraneans (i.e., Jewish, unpublished results).

However, technical problems of distinguishing DQS and DQ9 serologically should be taken into account to in-

terpret these results.

~26-~44. Only found in Algerians (HF: 1.5) and in

Portuguese (HF: 1.6) 1181, A26-B44 may represent an-

other Mediterranean haplotype.

A30-B18 (HF: 1.5). This haplotype is found in

Basques (HF: 3.7), Spaniards (HF: 3. l), and Sardinians (HF: 16.3) IlS]. It is a paleo-North African-Iberian

marker 13, 191.

~33~14 (HF: 2.5). This association is also found in

other Mediterraneans: Spaniards (HF: 2.8), Basques (HF: 1.2), Armenians (HF: 2.5), Portuguese (HF: 2.7), and Sardinians (HF: 3.0) [3, 181.


TABLE 4 HLA-DRB l-DRB3-DQA l-DQB 1 haplotypes with linkage disequilibrium in an Algerian population

Haplotype HF (%) LD X 100 RLD

DRB 1 0101 0102 1501 1501 1501 1502 1601 0301 0301 0302 0402 0402 0403 0403 0404 0404 0405 0406 1101 1101 1102 1102 1104 1201 1202 1301 1301 1301 1302 1302 1302 1303 1401 1401 1401

07 07 07

0801 0803 0804 0806 1001

DRB3 - - - - - - -

0101 0202 0101 - - - - - - - -

0202 0202 0202 0301 0202 0202 0202 0101 0202 0101 0301 0301 0301 0101 0201 0202 0202 - - - - - - - -

WA1 DQB 1 01 0501 01 0501

0102 0602 01 0602

0102 0502 0103 0601 0102 0502 0501 0201 0501 0201 040 1 0402

03 0302 03 0201 03 0301 03 0302 03 0302 03 0402 03 0302 03 0402

0501 0301 0102 0602 0501 0301 0102 0502 0501 0301 0501 0301 0601 0301 0103 0603 0103 0603 0103 0601 0102 0604 0102 0605

01 0604 0501 0301

01 0503 01 0503 01 0501

0201 0201 0201 0301 0201 0302 040 1 0402 0501 0301 0501 0301 0102 0602

01 0501

1.4” 1.4 1.00

8.0” 7.8 0.98 8.9’ 8.8 0.77 0.9” 0.8 0.07 1.4” 1.3 0.33 2.8’ 2.8 0.97 2.8’ 2.8 0.97 4.2’ 4.2 0.35 7.1b 6.9 0.57 1.4” 1.4 0.56 4.2’ 4.2 1.05 0.5” 0.4 0.10 0.5” 0.4 0.10 3.3b 3.2 0.80 0.3” 0.8 0.42 0.9” 0.9 0.47 2.4’ 2.4 1.00 1.9” 1.9 1.00 4.7b 4.6 0.92 0.5” 0.5 0.10 2.4’ 2.4 0.83 0.5” 0.5 0.17 2.8’ 2.8 0.97 0.5” 0.5 1.02 0.5” 0.5 1.02 1.9” 1.9 0.53 1.4” 1.4 0.39 0.5” 0.5 0.14 0.3” 0.9 0.25 0.5” 0.5 1.02 2.4b 2.4 0.67 2.4’ 2.4 0.96 0.5” 0.5 0.25 1.4” 1.4 0.70 0.5” 0.5 0.26 9.9b 9.7 0.82 0.5” 0.4 0.03 1.5” 1.4 0.12 0.9” 0.9 0.91 0.5” 0.5 1.06 0.Y 0.9 0.94 1.4” 1.4 0.93 3.3b 3.2 0.91

HF, haplotype frequency expressed as percentage; LD; Linkage disequilibrium obtained by the 1 lth HLA Workshop methodology {141; and RLD, relative linkage disequilibrium [ 151.

Alleles DQA 1’010 1 and 0 104 were all assigned as DQAI’O 1. Comparisons between DRB 1 *, DRB3*, DQAl*, and DQBl* alleles were made in 106 Algerians.

Standard deviation (SD) values are expressed as “0.004 < SD < 0.010 and ‘0.010 s SD s 0.025.

Results obtained were compared with chose in Refs. 131 and [ 181.

HLA-DRIDQ associations. alleles go with DRB 1” 160 1 in Caucasoids and also with

DRB lx 150 1 in Algerians (present article) and Spaniards

ASSOCIATIONS SO FAR ONLY FOUND IN ALGERIANS. [3, 261; (3) DRB1”1301-DRB3”0101-DQA1’0103- The associations so far not found in other populations DQBl”0601: these DQAl-Bl alleles go with { 181 and therefore considered characteristic markers for DRB 1” 1502 instead, in Caucasoids; (4) DRB l+ 1302- Algerians are (1) DRB 1’ 150 l-DQA l*O 10 l- DRB3’0301-DQAl”Ol-DQB1’0604; (5) DQB 1*0602; (2) DRB 1” 1102-DRB3*030 l- DRB1*0402-DQAl*03-DQB1*0201; (6) DQA 1’0 102-DQB l”O502: these particular DQAl-B 1 DRB 1”0406-DQA1’03-DQB 1+0402; (7) DRB 1*07-


TABLE 5 HLA-B, DR, DQ three-loci with linkage disequilibrium Algerian population

haplotypes in an

is included in the A30-B18-DR3 haplotype, a classical

Iberian and paleo-North African haplotype 119,291; (5)

DRB 1*0405-DQA 1*03-DQB 1*0302 is also described

in Spaniards 13 11 and considered a rarity in European and American Caucasoids {32-341, DRB l”O405 allele has

been found in other Mediterraneans such as Sardinians

IlSl, Ethiopians 1351, and Greeks [361 but associated with other DQ alleles; (6) DRB lx 140 l-DRB3*020 l-

DQAl*O l-DQB lx0503 is also described in Sardinians

E291; and (7) DRB 1*0402-DQA1*03-DQB 1*0302 is also found in high frequency in Israeli Jews 1371.

Haplotype HF (%) LD X 100

B-DRB I-DQA l-DQB 1 B 14-O 102-O l-050 1 3.4

B7-1501-0102-0602 1.9

B44-1502-0103-0601 0.9

B35-0301(DRB3’0101)-0501-0201 2.5

B IS-0301(DRB3+0202)-0501-020 1 2.5

B44-0402-03-0302 1.9

B49-1102(DRB3f0202)-050 L-0301 1.9

B35-1104(DRB3*0202)-0501-0301 1.5

B57-07-0201-0201 2.5

B57-07-020 l-0302 1.5

B38-0801-0401-0402 0.9

B7-1001-01-0501 1.9

3.2 1.6

0.9 1.9 2.2 1.8 1.8

1.5 2.2 1.2

0.9 1.8

HF. haplotype frequency expressed as percentage; LD, linkage disequilibrium obtained by the 1 lth HLA Workshop methodology [141.

Alleles DQAl*OlOl and 0104 were all assigned as DQAl*Ol.

Comparisons between HLA-B, DRBl+, DRB3*, DQAl*, and DQBl’ Al- leles were made in 102 Algerians.

Results obtained were compared with those in Refs. [31 and [181.

DQAl”020 l-DQB lx030 1; and (8) DRB 1*07- DQA l”O20 l-DQB 1’0302. These new-found associa-

tions were confirmed by automatic DNA sequencing (see

Materials and Methods).

ASSOCIATIONS FOUND IN ALGERIANS AND ALSO IN

OTHER POPULATIONS. Meditewaneans: (1) DRB 1*1502- DQA1*0102-DQB 1*0502 is also found in Spaniards

{26]; (2) DRB 1*0806-DQAl*O 102-DQB 1*0602 is the

most common DR8 subtype in Algerians [27] and is

found in Spaniards [3] and African Americans 1281; (3) DRB l+O 102-DQA lx0 l-DQB lx050 1 is present in high

frequency in Spaniards [3], Sardinians [29], Israeli Jews lb], and in North American Negroids (301; (4)

DRB 1*030 l-DRB3*0202-DQA lx050 1-DQB 1+020 1

FIGURE 2 Dendrogram showing relatedness between Al- gerians and other populations. Standard genetic distances were calculated by using HLA-A and -B (serologic) and -DRB 1, DQAl, and DQB 1 (DNA sequence) frequencies.

_ ALGERIANS

SVStlMEN

African Negroids: (8) DRB 1” 110 l-DRB3*0202- DQA1*0102-DQBl”0602 association is found in Ne-

groids 130, 381 and (9) DRB1#0404-DQA1”03- DQBl”0402 is also present in Bushmen and Hottentots

{18}. The latter DR/DQ association goes in favor of the

postulated Negroid component of Algerians 141.

Tentative Assignment of HLA Extended Haplotypes to the Different Original Ethnic Groups (Table 6)

Extended haplotypes in Algerians had been poorly studied previously; this was in part due to a lack of HLA

DNA sequencing technology E6, 39, 401. However, pre- vious data have been confirmed and new haplotypes have

been described (see Table 6). An exhaustive study has

been made in order to compare Algerian population with

Spaniards, Sardinians, Basques, and others E6, 18, 29,

38, 411. The data presently available will test the hypothesis of a common Iberian paleo-North African origin

of Spaniards, Basques, Sardinians, and Algerians 119,

27, 3 11. HLA data are accompanied with recorded his-

toric origins and migrations of North African, Iberian, and European populations.

The most representative haplotypes found in Algeri-

ans and their HFs are shown in Table 6. In addition, the following haplotypes have only been

found in Algerians: A2-Cblank-B38-DRS-

(DRB I”080 l)-DQ4(DQA lx040 l-DQB 1*0402) and

A2-Cw7-B7-DR lO(DRB lx 100 l)-DQ l(DQA 1”O l-

DQBl*0501) (HF: 0.9 and 0.9, respectively) { 18).

HLA Relatedness of Algerians with Spaniards and Basques: Paleo North Africans, Iberians, and Basques (Fig. 2)

In the last 1 lth HLA Workshop a dendrogram was constructed by using only HLA-A and -B frequencies and

the NJ method was done to calculate the genetic relationships among worldwide populations [25]. Spaniards, Basques, and Sardinians clustered together with African populations and were separated from the main European

branch of ethnic groups. Maghreb and Middle East populations could not be included in the analysis, because of the lack of data.

In the present work, not only HLA-A and -B frequen-


TABLE 6 The nine most frequent estimated extended haplotypes in Algerians and tentative assignment to the original contributing ethnic groups

HF Possible HLA-A,B,C HLA-DRB I-DRB3/4 HLA-DQAl-DQB 1 (%) origin

Al Cw7 B57 DR7(DRB1*07-DRB4*0101) DQ2(DQAl*O201-DQB1’0201)” 1.9 Iberian-paleo-North African (Basque-Algerian)

A30 Cw5 B18 DR3(DRB1*0301-DRB3+0202) DQ2(DQA1*0501-DQB1*0201)b 1.5 Iberian-paleo-North African

A33 Cw8 B14 DRl(DRBl*0102) DQ5(DQA1*Ol-DQBlf0501)r 1.5 Mediterranean

A2 cw4 B35 DRl1(DRBl’l104-DRB3’0202) DQ7(DQAIf0501-DQB1*0301)d 1.5 Mediterranean

A2 Cw4 B35 DR3(DRB1*0301-DRB3’0101) DC2(DQA1*0501-DQB1’0202)’ 1.5 Mediterranean (also in Sardinians)

A3 Cw7 B49 DRl l(DRB l* 1102.DRB3*0202) DQ7(DQA1*0501-DQB1*030ly 1.5 Autochthonous

A3 Cwl B7 DRlS(DRB1*1501) DQ6(DQAl*0102-DQBl*OGOZ)p 1.5 ?

A26 Cbl 844 DRlS(DRB1*1502) DQ6@QAlr0103-DQB1*0601)’ 0.9 Autochthonous

A2 Cw5 B44 DR4(DRB 1+0402-DRB4+0 10 1) DQ8(DQA1*03DQB1*0302) 0.9 Autochthonous

A2 Cw7 B7 DRlS(DRB1*1501) DQ6(DQA1*0102-DQBl*ObOZ)’ 0.9 ?

HF, haplotype frequency expressed as percentage. Alleles DQAl*OlOl and 0104 were all assigned as DQAl+Ol.

a Only common to French-Basques (HF: 2.7) and Spanish-Basques (HF: 1.2) I3.181. It is probably a haplotype of a paleo-North African-Iberian origin; it is different from the worldwide distributed Al-B57-DR7 haplotype which goes with Cw6 and DQ9. In a study done in 1981 in the AlgerianKabyle population, this haplotype was also found to be the second most frequent [39]. It was described as A I-Bw 17.

b Also found in Spaniards (HF: 3. I), Sardinians (HF: 11.4), and French-Basques (HF: 4.7) (18). Its frequency is much lower or absent in other Caucasoid populations. This haplotype was defined as Iberian-paleo-North African [ 19]. Presently, reported data on Algerians support this hypothesis. Bouali et al. 1391 found this haplotype to be the first most frequent in Algerians; it was described as B18-BfFl.

’ Shows the highest frequency in Armenians (HF: 3.1) and is also high in frequency or present in other Medirerranean Caucasoid populations and as Sardinians (HF: 2.7). Spaniards (HF: 2.6). French (HF: 1.4). Greeks (HF: 1. l), and Italians (HF: 0.7) [ 181. Bouali et al. [39] found it in Algerians. This haplotype does not appear in Basques and has a much lower frequency in northern European Caucasoid populations [3]. The fact that Armenians (and not Basques) bear this marker does not support a theory that Basques and Armenians are related and with common descent [3]. Armenians may have acquired this haplotype during their historical Middle East Mediterranean setrlements in the middle ages [451. Alternatively, this haplotype may be of Armenian origin.

’ Also found in Italians (HF: 2.3) and Greeks (HF: 2.0) [18]. B35-DRI l-DQ7 is also present in non-lberian Mediterranean areas such as Albania (HF: 3.3), Armenia (HF: 5.0). and France (HF: 1.1) and in other populations such as Germans (HF: 3.4), Indians (HF: 3.1). and Hungarians (HF: 3.9) [ 181.

’ Common to Sardinians (HF: 2.7) and French (HF: 1.0) (181. HLA-B35-DR3-DQ2 is also present in Greeks (HF: 1.9) and Uralic populations (HF: 3.1) [18]. It could be a non-Iberian Mediterranean haplotype.

’ Only present in Algerians.

8 Appears in Spaniards (HF: 1.4) [3] and has been defined as a northern European haplotype [29] due to its high frequency in Danes (HF: 3.6). Austrians (HF: 3.2), Czecks (HF: 2.5). Germans (HF: 2.5), Yugoslavians (HF: 2.4). and also in French (HF: 1.8) and Italians (HF: 0.6) [18].

b Common to French-Basques (HF: 3.6). Cornish (HF: 3.6). Austrians (HF: 2.6). Britons (HF: 2.3), and Spaniards (HF: 1.9) I3,18] and does not appear in high frequency in northern Cxmssoids, such as Scandinavian, or in other Mediterraneans. This haplotype is extended throughout England, Spain, and northern Africa.

ties, but also DR ones, have been used to test relatedness between Algerians and other Europeans, particularly

northern Mediterraneans such as Spaniards, Basques, and

Sardinians. The same 1 lth Workshop methodology was

used (see Materials and Methods). Our results fully con-

firm the 11th Workshop data; Basques and Spaniards

cluster together with an African population, Algerians.

This also supports the theory that Iberians and paleo- North African Hamites (mainly represented by Berbers at present) are genetically related populations with com-

mon ancestors and that Basques may be a relative isolate of these primitive Iberian populations 13, 191. The ge-

netic system used (HLA-A and -B serology and DRB 1,

DQAl, DQB 1 exon-2 DNA sequences) is one of the most powerful and polymorphic genetic systems which may be utilized to discriminate populations (see above). In addition, results obtained analyzing haplotype frequencies, particularly of the characteristic haplotype A30-B18 and Al-B57-Cw7, which are probably markers

of neolithic Iberian people (Table 6) {31, further support a relatedness between paleo-North Africans and present- day Spaniards and Basques (Iberians). Other genetic data that apparently identified Basques from surrounding populations were mostly based on blood groups. Rh-

negative phenotype high frequency and B group low

frequency, once regarded as distinctive of Basque populations [42], has been found to be a general characteristic

of other neighboring western European populations. Rh-

negative gene frequencies (cde]: Spanish-Basques (0.44),

Portuguese (0.54), Isle of Man (0.43), Spaniards (0.38),

French (0.41), Irish (0.42), and English (0.39). B-gene frequencies: Basques (0.03), Portuguese (0.05), Irish

(0.07), Spaniards (O.O6), French (O.O6), and English (0.06) 151. Thus frequencies of Basque Rh-negative and B phenotype are not significantly different from sur-

rounding populations, although a distinctive effect

might be expected due to the historic Basque relative

lack of substantial admixture with neighbors and invad-

ers 143, 441. In summary, in the present work complete HLA data

on Basques and Spaniards was compared for the first time with similar HLA data on a North African population (Algerian) with a high paleo-North African (Berber)

component [4, 271. Figure 2 (dendrogram) shows that Basques and Spaniards are closer to Algerians than to any other Caucasoid, Negroid, or Mongoloid population. In addition, the specific highly discriminative HLA haplotype, HLA-A30-B 1%DR3, is a marker common to Sar-

HLA in Algerians

dinians, Basques, Spaniards, and Algerians. Taken together, these data are the strongest reported genetic support for a common descent of Basques, Spaniards, and Algerians, probably reflecting the preneolithic relationship between Iberians and paleo-North Africans.

ACKNOWLEDGMENTS

This work was supported by Fondo de Investigaciones Sani-

tar&, Ministerio de Sanidad, and Fundacion Ramon Areces, Spain. We are grateful to Manuel Martin-Villa, Foutia

Bekhoucha, and Fatima Belhani for their collaboration. We

thank to Dr. Nei, Dr. Kumar, and Dr. Ota for the computer programs kindly provided.

REFERENCES

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

McEvedy C, Jones R: Atlas of World Population History.

Harmondsworth, Penguin Books, 1978.

Bodmer WF, Cavalli-Sforza LL: Genetics, Evolution and

Man. San Francisco, WH Freeman 1976, p 573.

Martinez-Las0 J, De Juan D, Martinet-Quiles N, Gomez- Casado E, Cuadrado E, Arnaiz-Villena A: The contribu-

tion of the HLA-A, -B, -C and -DR, -DQ DNA typing to the study of Spaniards and Basque origins. Tissue An-

tigens 45:237.

Julien CA: Historie de I’Afrique du Nord. Paris,

PAYOT, 196 1.

Roychouldhury AK, Nei M: Human Polymorphic Genes World Distribution. New York, Oxford University

Press, 1988.

Brautbar C, Friedman A, Battat S, Tambur AR, Cam-

bon-Thomsen A, Ohayon E, Tongio M, DuPont E, Es- sakali-Gharbi M: HLA in Israeli Jew, Moroccans and Al-

gerians. In Tsuji K, Aizawa M, Sasatuki T (eds). HLA

1991, vol 1 Oxford, Oxford University press, 1992, p

656.

Danilovs J, Terasaki PI, Park MS, Ayoub G: B lympho- cyte isolation by thrombin-nylon wool. In Terasaki PI

(ed): Histocompatibility Testing 1980. Los Angeles, UCLA Tissue Typing Laboratory, 1980, p 287.

Davis RW, Thomas M, Cameron J, St John TP, Padgett RA: Rapid isolation for enzymatic and hybridization analysis. Methods Enzymol 65:404, 1980.

Core11 A, Martin-Villa JM, Varela P, Vicario JL, Mar- tinez-Laso J, Arnaiz-Villena A: Exon 2 DNA sequence of the DRw 13b allele obtained from genomes of five different individuals. Mol Immunol 27:313, 1990.

Kimura A, Sasazuki T: Eleventh International Histocom- patibility Workshop Reference Protocol for the HLA-

DNA-Typing Technique. In Tsuji K, Aizawa M, Sasa- zuki T (eds): HLA 199 1, vol 1. Oxford, Oxford Univer- sity Press, 1992, p 397.

Nepom GT, Byers P, Seyfried C, Healey LA, Wilske KR, Stage D, Nepom BS: HLA genes associated to rheu-

12. Gocayne J, Robinson DA, Fitzgerald MG, Chung FZ, Kerlavage AR, Lentes KU, Lai J, Wang CD, Fraser CM, Venter FG: Primary structure at rat cardiac P-adrenergic and muscarinic cholinergic receptors obtained by automated DNA sequence analysis: further evidence for a

multigene family. Proc Nat1 Acad Sci USA 84:8296,

1987.

13.

14.

Mattiuz PL, Ihde D, Piazza A, Ceppelini R, Bodmer WF:

New approaches to the population genetics and segrega- tion analysis of the HLA system. In Histocompatibility

Testing 1970. Copenhagen, Munskaard, 1970, p 193.

Imanishi T, Akaza T, Kimura A, Tokunaga K, Gojorobi T: Estimation of allele and haplotype frequencies for HLA and complement loci. In Tsuji K, Aizawa M, Sasazuki T

(eds): HLA 199 1, vol 1. Oxford, Oxford University Press,

1992, p 76.

15.

16.

Degos L: La repartition anthropologique des genes HLA et dynamique des populations. In Dausset (ed): HLA

1982. Complexe majeur d’histocompatibilite de I’homme. Flammarian Medicine Sciences, 1982, p 13 1.

Pickbourne P, Piazza A, Bodmer WF: Population analy-

sis. In Bodmer WF, Batchelor JR, Bodmer JG, Festesn- tein H, Morris PJ (eds): Histocompatibility Testing

1977. Copenhagen, Munksgaard, 1978, p 259.

17. Vicario JL, Martinez-Laso J, Core11 A, Regueiro JR, Ar- naiz-Villena A: Both HLA class II and class III restriction fragment length polymorphism factors linked to rheuma- toid juvenile arthritis. Clin Immunol Immunopathol 56:

22, 1990.

18. Imanishi T, Akaza T, Kimura A, Tokunaga K, Gojorobi

T: Allele and haplotype frequencies for HLA and complement loci in various ethnic groups. In Tsuji K, Aizawa M, Sasazuki T (eds): HLA 1991, vol 1. Oxford, Oxford

University Press, 1992, p 1065.

19. Arnaiz-Villena A, Rodriguez de Cordoba S, Vela F, Pas- cual JC, Cervero J, Bootello A: HLA antigens in a sample

of the Spanish population: common features among Span-

iards, Basques and Sardinians. Hum Genet 58:344, 1981.

20. Saitou N, Nei M: The neighbor-joining method: a new method for reconstructing phylogenetic tree. Mol Biol Evol 4:406, 1987.

21. Nei M: Genetic distances between populations. Am Nat 106:283, 1972.

22.

23.

Rose NR, Macario EC, Fahey JL, Friedman H, Pen GM (eds): Manual of Clinical Laboratory Immunology. Wash-

ington, American Society for Microbiology, 1992, p 901.

Snell GD, Dausset J, Nathenson S (eds): Histocompati- bility: HLA Polymorphism. New York, Academy Press, 1976, p 210.

24. Tsuji K, Aizawa M, Sasazuki T (eds): HLA 1991: No-

267

matoid arthritis: identification of susceptibility alleles us-

ing specific oligonucleotide probes. Arthritis Rheum 32:

15, 1989.

268 A. Arnaiz-Villena et al.

25.

26.

27.

28.

29.

30.

31.

32.

33.

menclature for Factors of the HLA System, 1991, vol 1. oxford, Oxford University Press, 1992, p 17.

Imanishi T, Wakisaka A, Gojorobi T: Genetic relation-

ships among various human populations indicated by MHC polymorphisms. In Tsuji K, Aizawa M, Sasazuki T


1992, p 627.

Vilches C, De Pablo R, Moreno M, Solis R, Kreisler M: Characterization of an HLA-DR15 DQS haplotype found

in the Spanish Caucasoid population. Hum Immunol 35:

223, 1992.

Benmamar D, Martinez-Las0 J, Varela P, Bekhoucha F, Morales P, Arnaiz-Villena A: Evolutionary relationships

of HLA DR8 alleles and description of a new subtype (DRB1*0806) in the Algerian population. Hum Immu-

no1 36:172, 1993.

Apple RJ, Erlich HA: Two new HLA-DRB 1 alleles found

in African Americans: implications for balancing selection

at positions 57 and 86. Tissue Antigens 40:69, 1992.

Piazza A, Fauchet R, Richiardi P, Carcassi C, Contu L:

Anthropology report: Belgian, French, Italian, Portu-

guese, Sardinian and Spanish population. In Tsuji K,

Aizawa M, Sasazuki T (eds): HLA 1991, vol 1. Oxford,

Oxford University Press, 1992, p 648.

Fernandez-Viiia MA, Gao X, Moraes ME, Moraes JR,

Salatiel I, Miller S, Tsai J, Sun Y, An J, Layrisse Z, Gazit E, Brautbar C, Stastny P: Alleles at four HLA class II loci determined by oligonucleotide hybridization and their as-

sociations in five ethnic groups. Immunogenetics 34:299,

1991.

Morales P, Martinez-Las0 J, Martin-Villa JM, Core11 A, Vicario JL, Varela P, Arnaiz-Villena A: High frequency of

the DRB 1*0405-(Dw 15)-DQw8 haplotype in Spaniards

and its relationship to diabetes susceptibility. Hum Im-

munol 32: 170, 199 1.

Lanchbury JSS, Hall MA, Welsh KI, Panayi GS: Se-

quence analysis of HLA-DR4B 1 subtypes: additional first domain variability is detected by oligonucleotide hybrid-

ization and nucleotide sequencing. Hum Immunol 27:

136, 1990.

Gao XJ, Fernandez-Vina M, Shumway W, Stastny P: DNA typing for class II HLA antigens with allele specific

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

or group specific amplification: typing for subsets of DR4. Hum Immunol 27:40, 1990.

Celis E, Karr RW, Gregersen PK, Reismoen NL: Func- tional analysis of MHC class II-restricted T cells derived

from a Caucasian with a DR4, Dw15, DQw8 haplotype. Hum Immunol 29:31, 1990.

Stastny P, Kimura A: DNA typing analysis of the DR4 group. In Tsuji K, Aizawa M, Sasazuki T (eds): HLA

1991, vol 1. Oxford, Oxford University Press, 1992, p 465.

Cutbush SD, Sachas JA, Papasteriades C, Boki K, Biro PA: HLA associations in ethnic populations: HLA-Dw 15

is associated with HLA-DR4 in Greek population. Pro- ceedings of the British Society for Immunology, Spring

Meeting, London, April 10-12, 1991, p 41.

Cohen N, Mickelson E, Amar A, Battat S, Hansen JA, Brautbar C: HLA-Dw clusters associated with DR4 in Israeli Jews and the definition of a new DR4 associated

Dw subtype: Dw ‘SHA.’ Hum Immunol 19:179, 1987.

Stastny P, Kalil J: HLA in North American and South

American Negroids. In Tsuji K, Aizawa M, Sasatuki T


1992, p 638.

Bouali M, Dehay C, Benajam A, Poirier JC, Degos L, Marcelli-Barge A: HLA-A, B, C, Bf and glyoxalase I polymorphisms in a sample of the Kabyle population (Al-

geria). Tissue Antigens 17:501, 1981.

Mercier P, Vallo JJ, Vialettes B, Vague Ph: HLA-A, B, DR antigens and insulin-dependent diabetes in Algeri-

ans. Tissue Antigens 26:20, 1985.

Lindblom B, Svejaard A: HLA genes and haplotypes in the Scandinavian populations. In Tsuji K, Aizawa M,

Sasazuki T (eds): HLA 199 1, vol 1. Oxford, Oxford Uni-

versity Press, 1992, p 65 1.

Mourant AE: The blood groups of the Basques. Nature

160:505, 1947.

Caro Baroja J: Los Pueblos de Espaiia, vol 2. Madrid,

Ediciones Itsmo, 1981, p 33.

Roman de1 Cerro JL: El Origen Iberico de la Lengua Vasca. Alicante, Editorial Aguaclara, 1993, p 33.

Duby G: Atlas Historique. Librairie Larousse, 1987.

HLA allele and haplotype frequencies in Algerians

Documents