Download β2-microglobulin gene mutation is not a common mechanism of HLA

© Springer-Verlag 2000
Int J Clin Lab Res (2000) 30:87–92
ORIGINAL
M.A. Fernández • F. Ruiz-Cabello
M.R. Oliva • T. Cabrera • P. Jimenez
M.A. López Nevot • F. Garrido
β2-microglobulin gene mutation is not a common mechanism
of HLA class I total loss in human tumors
Received: 15 February 2000/ Accepted: 8 March 2000
Abstract One hundred and sixty-two tumor samples were
analyzed for HLA class I expression using immunohistological techniques. HLA class I total loss (phenotype no. I) was
detected in 31 cases (19%), comprising 20 colorectal, 3
laryngeal, and 2 bladder carcinomas and 6 melanomas.
Twenty-one cases were selected for molecular analysis due
to a higher proportion of tumor cells versus stroma cells
(75%). We investigated whether β2-microglobulin mutation
was responsible for HLA downregulation. Single-strand
conformation polymorphism and sequencing analysis of
DNA samples was performed. Alterations were detected
only in melanomas M78 (a point mutation in the initiation
ATG sequence), M79 (a mutation in codon 31 producing a
stop codon), and M34 (a TTCT deletion introducing a termination codon signal). We found no β2-microglobulin gene
mutation in the other 18 samples. Loss of heterozygosity in
15q close to the β2-microglobulin gene was found in 5 cases.
We conclude that HLA class I total loss can frequently occur
without β2-microglobulin gene mutations.
Key words β2-microglobulin • Gene mutation • HLA class I
total loss • Human tumors
M.A. Fernández • F. Ruiz-Cabello • M.R. Oliva • T. Cabrera
P. Jimenez • M.A. López Nevot • F. Garrido
Servicio de Análisis Clínicos, Hospital Universitario Virgen
de las Nieves, Universidad de Granada, Avda.
de las Fuerzas Armadas s/n, 18014 Granada, Spain
F. Garrido ()
Servicio de Análisis Clínicos, Hospital Universitario Virgen de las
Nieves, Avda. de las Fuerzas Armadas s/n, 18014 Granada, Spain
Introduction
Major histocompatibility (MHC) class I molecules play an
important role in cell recognition against virus-infected and
tumor cells. Their main function is to present peptides
derived from intracellular proteins to cytotoxic T-lymphocytes (CTL). The correct assembly of MHC class I peptide
complexes is required for the stable expression of HLA class
I [l]. This assembly occurs in the endoplasmic reticulum,
where endogenous β2-microglobulin (β2m) gene expression
is necessary for proper intracellular MHC assembly and stabilization by endogenous peptides. Tumor cells frequently
lose HLA molecules during tumor development [2]. These
alterations in HLA gene expression can provide the tumor
cells with a similar mechanism of escape from the immune
system [3] to that used by some viruses [4]. Tumor cells
often exhibit a complete loss of expression of HLA antigens;
this is a relatively frequent phenotype (9%–52%), associated
with mutations in the β2m gene or defects in the transporterassociated processing. The negative impact of β2m gene
mutations on T-cell based immunotherapy has been reported
[5, 6], indicating the importance of analyzing the HLA phenotype of the tumor.
The aim of this study was to investigate the frequency of
β2m mutations in a series of melanomas and fresh head and
neck, bladder, and colorectal cancers previously studied for
HLA class I antigen expression. Our results indicate that
somatic β2m gene mutation in human tumors showing total
loss of HLA expression is not a frequent event.
Materials and Methods
Patients and tumor specimens
The present study was performed on samples from patients diagnosed with melanoma, head and neck, bladder, and colon carcino-
M.A. Fernández et al.: β2m mutation in human tumors
88
ma. Samples of tumor tissues were provided by different departments of the Virgen de las Nieves University Hospital in Granada,
Spain. Tumor tissues were snap-frozen in isopentane cooled with
liquid nitrogen within 1–2 h of removal, and were stored in liquid
nitrogen prior to study. Cryostat sections 5–8 µm thick were cut
(Bright microtome), allowed to dry at room temperature for 4–18 h,
fixed for 10 min in acetone, and stored at –40ºC prior to staining.
Exon 2:
β2C5´: 5´ CGA TAT TCC TCA GGT ACT 3´
β2NC3´: 5´ CAA-CTT-TCA-GCA-GCT-TAC 3´
Cell lines
with annealing temperatures of 56º and 52ºC, respectively.
Amplified samples (2.5 µl) were mixed with 9 µl of sequencing
stop solution (USB), 1.5 µl of 0.08 N sodium hydroxide, and 1.5 µl
of 0.1% sodium dodecyl sulfate; they were denatured for 10 min at
95ºC and rapidly cooled in dry ice. Samples (3 µl) were loaded onto
an 8% non-denaturing acrylamide gel containing 10% glycerol and
run at room temperature for 18-20 h at 6-8 W. The gels were dried
at 80ºC under vacuum and exposed to X-ray film for 20-24 h.
LoVo and HCT-15 colorectal tumor cell lines were grown at 37ºC
in a humidified atmosphere of 5% CO2 in RPMI medium with 10%
fetal bovine serum.
DNA sequencing
Immunohistological staining and HLA tumor typing
Tumor sections were stained by a biotin-streptavidin amplified
detection system (supersensitive multilink-HRP/DAB, BioGenex),
as previously described [7]. Monoclonal antibodies used were:
W6/32 against HLA class I heavy chain/β2m complex, GRH1
against free β2m, and HC-10 against the α chain of HLA-B-C molecules [8, 9].
DNA extraction
Genomic DNA was extracted from cell lines, tumors, and normal
tissues (mucosa or blood). Frozen solid tissues were ground using
a mortar and pestle with liquid nitrogen. DNA extraction buffer [5
ml/g of tissue) containing 10 mM TRIS-HCl pH 7.5, 150 mM
NaCl, and 2 mM EDTA pH 8.0 was added to the homogenate.
Sodium dodecyl sulfate at a final concentration of 0.5% and proteinase K (200 µg/ml) were added to the resulting suspension and
the mixture was incubated overnight at 37ºC. After extraction with
phenol-chloroform-isoamylalcohol (25:24:1) (v:v:v), the DNA
was precipitated in ethanol, dried and resuspended in TRIS-EDTA.
The same protocol was used for blood samples after red cells lysis
and for cell lines.
Single-strand conformation polymorphism assays
Single-strand conformation polymorphism (SSCP) of normal,
tumor, and cell line DNAs was performed according to a standard
method [10] with slight modifications [11]. Polymerase chain reaction (PCR) amplifications were performed using 100 ng of genomic DNA in 10-µl volumes. The sequences of the primers used were
obtained from intronic regions flanking leader peptide/exon 1 and
exon 2 of the β2m gene [12]:
Exon 1:
β215´: 5´ CTG-ATT-GGC-TGG-GCA-CGC 3´
β213´: 5´ TGA-GAA-GGA-AGT-CAC-GGA-GC 3´
Genomic DNA (500 ng) for direct sequencing was amplified separately from the amplified material used for SSCP, but using the
same annealing temperatures and primers (250 ng) as above. PCR
products for exon 1 and exon 2 were 237 and 365 base pairs respectively. The amplicons were sequenced with Big Dye Terminator
Cycle Sequencing Kit and the ABI-PRISM 377 DN Sequencer (PE
Applied Biosystem). Sequencing was carried out in both directions
using the same primers as those used for PCR amplification. The
sequencing reaction contained 30 ng of PCR product, 8 µl of premix, and 3.2 picomoles of forward and backward primers, in a total
of 200 µl. This mixture was heated to 94ºC for 3 min, followed by
incubation for 25 cycles (96ºC 10 s, 52ºC 15s, 60ºC 1 min). The
final products were purified with Centri-sep columns (Pricenton
Separation), denatured in polyacrylamide gel, and analyzed with
DNA sequencing analysis software (PE Applied Biosystem).
Microsatellite analysis
Loss of heterozygosity (LOH) for the β2m gene at 15q21 was
assessed with two highly polymorphic dinucleotide repeat markers.
The markers were chosen on the basis of their heterozygosity and
location. The microsatellite markers studied on chromosome 15
were D15S126 and D15S165 [13], which flank the β2m gene. We
used a fluorescent microsatellite assay, and PCR reactions were performed in a total volume of 15 µl, containing 60 ng of each DNA
sample, 1× PCR buffer, 5 µM each of unlabelled primer, and 5’ endlabelled primers with fluorescent dyes, 0.5 units Taq DNA polymerase, and 250 µM of each dideoxynucleotide. A portion of the PCR
products was aliquoted and combined with dextran blue dye, formamide, and GeneScan TAMRA internal size marker to permit precise sizing of alleles. Samples were denatured at 94ºC for 2 min,
snap-cooled on ice, and loaded onto a 6% denaturing polyacrylamide
gel containing 7.7 M urea, and analyzed on an ABI 377 automated
sequencer. Each PCR was carried out as a “single” reaction and then
pooled for electrophoresis. Specific Genescan and Genotyper software (PE Applied Biosystem) was used to size, quantify, and compare normal and tumor amplicon patterns for marker.
LOH was calculated by determining a ratio of the two allele
pairs correct for differences in the amplification efficiencies of the
normal/tumor DNA samples. LOH was considered if the reduction
rate of the height of the allele in the tumor was more than 25% [14].
M.A. Fernández et al.: β2m mutation in human tumors
Results
Expression of HLA class I antigen in primary tumors
A total of 162 samples of fresh human tumors (32 head and
neck, 5 bladder, and 88 colorectal carcinomas and 37
melanomas) were studied for HLA class I expression. Frozen
sections from each tumor sample were analyzed for HLA ABC
total losses (phenotype I). This phenotype was established when
monoclonal antibodies against heavy chain and β2m were
simultaneously negative in a significant portion of the section
(according to the classification of Garrido et al. [2]). The positive controls were the intrinsic stromal cells in each section
(endothelium, infiltrating lymphocytes, and fibroblasts).
Table 1 shows the results for total HLA losses in the
tumors included in the study and does not reflect allelic or
Table 1 Frequency of total HLA-ABC lossesa
Tumour type
HLA-ABC
Loss (%)
Colon
Larynx
Melanoma
Bladder
23
9
16
40
aA
89
locus-specific losses, which were not analyzed. A total of 31
cases (19%) showed HLA class I total losses, a proportion
that is consistent with preliminary observations for this phenotype. Twenty-one cases of HLA class I total loss were
selected for molecular analysis due to a higher proportion of
tumor cells than stroma cells (>75%). In 19 cases the loss of
expression was homogeneous, and in another 2 the pattern of
HLA expression was heterogeneous (Table 2). Heavy chain
intracytoplasmic expression was studied with monoclonal
antibody HC-10 and only 2 tumors were positive, indicating
that most of the β2m-negative tumors did not have detectable
free heavy chain in the cytoplasm.
Figure 1 depicts two representative samples showing
phenotype no. I total loss of HLA class I expression. Tumor
cells from the CO5 sample showed a negative pattern with
the W6/32 monoclonal antibody (Fig.1A). However, staining
of the tumor cells with an antibody that recognizes HLA
class I free heavy chain (HC10) (Fig.1B) revealed an accumulation of free heavy chain in the cytoplasm. In contrast,
the CO85 tumor cells showed a negative pattern with both
monoclonal antibodies (Fig. 1C and D).
The β2m gene in tumors with total loss of HLA expression
Twenty-one DNA samples from tumors with defective β2m
protein expression were examined for mutations in exons 1
total of 162 tumor samples were studied
Table 2 Study of β2-microglobulin (β2m) in HLA-ABC-negative tumors (phenotype I) (LOH loss of heterozygosity, del deletion, CO colorectal carcinoma, CL laryngeal carcinoma, VE bladder carcinoma, M melanoma)
Tumora
β2m protein
Free HLA heavy chain
β2m gene mutation
LOH in β2m region
CO5
CO14
CO18
CO19
CO22
CO26
CO40
CO44
CO85
CO86
CL56
CL79
CL80
VE4
VE6
M6
M24
M44
M78
M79
M34
–
–
–
–
–
–
+/+/–
–
–
–
–
–
–
–
–
–
–
–
–
+
–
–
–
–
–
–
–
–
+
–
–
–
–
–
–
–
–
–
–
–
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
T—>A
C—>G
TTCT del
LOH
No
LOH
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
LOH
LOH
LOH
a Tumor
samples were primary lesions with the exception of melanomas
M.A. Fernández et al.: β2m mutation in human tumors
90
Fig. 1 Immunohistological staining
of cryostat sections from patients
CO5 (A, B) and CO85 (C, D). In
CO85, tumor cells were not
stained with W6/32 or HC10 monoclonal antibodies. In CO5, accumulation of cytoplasmic heavy
chain was detected with monoclonal antibody HC-10 (B). W6/32 was
negative on tumor cells (A)
A
B
C
D
and 2 of the β2m gene by SSCP and sequencing. For the
SSCP assays, PCR products were labelled with 32P and analyzed under standard conditions. None of the cases studied
showed an abnormal pattern of SSCP in exons 1 and 2 of the
β2m gene. DNA from Lovo (CT deletion) and HCT-15
(mutated in the codon 10) cell lines were used as positive
controls in the SSCP assays. Figure 2 illustrates the SSCP
patterns of a set of the analysed tumor DNAs. A wild-type
band pattern (NT) was found in the DNA of tumors and a
band-shift pattern was seen in the DNA of Lovo, suggesting
that the fresh tumors had no mutations in the fragment analyzed. These results were confirmed in the sequencing exper-
LoVo NT
5 14 18
19 22
26 40
44
iments (Fig. 3). We examined the leader peptide sequence/exon 1 and exon 2 of the β2m gene for mutations. The
sequencing data revealed that HLA class I (heavy chain and
β2m) downregulation was due in most cases to other mechanisms and not to β2m mutations. The exceptions were 3
HLA-ABC-negative melanomas (Table 2). M78 and M79
were obtained from patients included in clinical trials using
MAGE-1-MAGE-3-HLA-A1 peptides [6]. In both cases, the
lack of HLA class I expression could be explained by specific mutations of the β2m gene. One, M78, was a T A
point mutation in the initiator ATG sequence, and the other
case was a C G point mutation in the codon 31, resulting in
85 86
Fig. 2 Single-strand conformation polymorphism pattern of human
β2-microglobulin (β2m) gene exon 1 polymerase chain reaction
(PCR) products from samples of HLA-negative colon carcinomas.
A band shift was observed in the colon carcinoma cell line Lovo
that contains a CT deletion. The pattern obtained with DNA from
these tumors was identical to the negative control (NT)
Fig. 3 Sequence data of β2m exon 2 PCR products. HCT-15 colorectal cell line showed a heterozygous mutation C A (arrow).
An HLA-negative colon carcinoma (CO-85) showed a normal
sequence in the β2m gene
M.A. Fernández et al.: β2m mutation in human tumors
Fig. 4 Loss of microsatellite D15S126 in the M34 melanoma tumor
cell line. PCR products of DNA from peripheral blood lymphocytes were compared with the corresponding DNA from
melanoma. The tumor shows the loss of a large allele (right)
a UGA stop codon in the second exon. M34 presented a
frameshift due to a TTCT deletion at an eight-base pair CT
repeat region of exon 1. This mutation introduced a termination codon signal in position 48.
Because these tumors apparently contained a β2m mutation affecting a single allele, we then asked whether the
remaining wild-type allele was lost by gene deletion. We used
two microsatellite markers flanking the β2m gene to determine allelic losses compatible with the loss of a full β2m gene.
The HLA-ABC-negative melanomas (M78, M79, and M34)
showed β2m mutations and a pattern of LOH compatible with
the deletion of the second allele (Table 2, Fig. 4). Two additional samples (CO5 and CO 18) exhibited LOH patterns and
no mutations in the β2m gene, indicating that an additional
mechanism must be implicated in HLA-ABC total loss.
Discussion
Several mechanisms have been implicated in the generation
of total or selective HLA-ABC losses [1, 2]. The complete
absence of HLA-ABC may be related to mutations in the
β2m gene or defects in the transporter-associated antigen
processing. Mutations in the β2m gene in human colon tissues and tumor cell lines have been related to a mismatch
repair defective phenotype [12, 15], because the β2m gene
contains a repetitive (CT)4 sequence in exon 1 and two (A)5
repetitive sequences in exon 2. However, the mutator phenotype is not a widespread characteristic of sporadic human
tumors, so that the frequency of β2m gene mutations may be
different according to the histological tumor type. In fact,
published data from our laboratory showed strong differences in the frequency of tumors with complete absence of
HLA-ABC molecules [2]. This prompted us to investigate
the frequency of somatic β2m gene mutations in 21 of 31
91
human tumors of different origin that exhibited phenotype
no. I (total loss of HLA expression). The tumor samples were
checked with hematoxylin staining for tumor tissue and
DNA was derived from specimens containing at least 75% of
tumor cells. The study combined immunohistological analysis of HLA-ABC molecules with a molecular approach to
investigate β2m gene mutations, to explore whether this
alteration is present in the majority of tumor cells.
Our group and others reported the negative impact of
β2m gene mutations on the T-cell-based immunotherapy of
melanoma [5, 6]. β2m expression is essential for cells to
form a functional HLA complex, and the mutation of the
β2m protein probably prevented the recognition of these
cells by specific CTLs in these patients (M78, M79).
However, our data from a series of sporadic HLA-negative
tumors of different histological types do not support the
proposition that β2m structural defects are a common mechanism for HLA-ABC total loss. We analyzed the DNA from
the tumor samples for the detection of mutations by SSCP
and sequencing experiments. Mutations were detected in
50% of the melanoma specimens studied (3 of 6 HLA- ABCnegative melanomas), but not in HLA-ABC-defective
tumors of other histological origin (Tables 1 and 2). We
could not confirm that β2m gene mutations are more frequent in colorectal cancer [15]. However, these mutations
may be a common feature of HLA-negative melanomas [5,
16, 17]. Because microsatellite instability is not frequently
observed in melanoma cells, these mutations do not appear
to be closely linked to a defective mismatch repair phenotype
[18]. In the 3 present melanomas that contained β2m gene
mutations, the other β2m gene was deleted by somatic LOH
and therefore presented HLA class-I negative phenotypes. In
this context, recent data from our laboratory (Cabrera et al.,
submitted for publication) indicate that LOH associated with
the β2m gene is a frequent finding in HLA-deficient tumors
that, however, may express some HLA alleles. This genetic
defect may be an early event happening before a β2m mutation
hit the other homologous gene. This β2m LOH is detected in
DNA obtained after tumor tissue microdissection to avoid contaminating stroma. We believe that the figures of β2m LOH
presented in Table 2 are therefore an underestimate, since
extraction of total tumor stroma DNA could mask the LOH.
The accumulation of defects in antigen processing and
presentation may also generate a phenotype characterized by
low expression or complete absence of HLA cell surface
expression. In fact, loss of TAP-1 expression has also been
found in a high percentage of human cervical cell carcinomas [19, 20]. However, the simultaneous absence of both
free heavy chain and β2m protein may indicate that a regulatory mechanism could be present in many tumors with low
levels of HLA class I expression. MHC class I and β2m
genes share conserved regulatory elements that control the
constitutive and inducible synchronous expression of these
proteins [21, 22]. Low transcriptional factor binding activity
has been demonstrated in human neuronal cells and other
M.A. Fernández et al.: β2m mutation in human tumors
92
HLA-deficient tumor cell lines [23, 24]. Further studies will
define whether this is an important mechanism in the majority of HLA-negative tumors with no β2m mutations.
The present data indicate that, except in the melanomas,
somatic β2m gene mutations were not responsible for the
total loss of HLA expression. In this context, there is a recent
report on head and neck carcinomas indicating that mutation
in the β2m gene is not a frequent event [25]. Understanding
of the exact mechanism by which cancer cells can elude the
immune response (regulatory versus structural defects) may
aid decision-making on vaccination strategy and therapeutic
approach.
Acknowledgements We would like to thank Miss Carmen
Amezcua, Immaculada Delgado, and Josefa Gil for expert technical
assistance. This work was supported by the Fondo de
Investigaciones Sanitarias, Consejeria de Salud (SAS) and Plan
Andaluz de Investigación, Spain.
References
1. Parmer E, Cresswell P. Mechanisms of MHC class I restricted
antigen processing. Ann Rev Immunol 1988; 16:323.
2. Garrido F, Ruiz-Cabello F, Cabrera T, Pérez-Villar JJ, LópezBotet M, Duggan-Keen M, Stern P. Implications for immunosurveillance of altered HLA class I phenotypes in human
tumors. Immunol Today 1997; 18:89.
3. Ruiz-Cabello F, Garrido F. HLA and cancer: from research to
clinical impact. Immunol Today 1998; 19:539.
4. Ploegh HL. Viral strategies of immune evasion. Science 1998;
280:248.
5. Restifo NP, Marincola FM, Kawakami Y, Taubenberger J,
Yannelli JR, Rosenberg SA. Loss of functional beta2microglobulin in metastatic melanomas from five patients
receiving immunotherapy. J Natl Cancer Inst 1996; 88:100.
6. Benitez R, Godelaine D, Lopez Nevot M, Brasseur F,
Marchand M, Cabrera T, Baren N van, Andry G, Jimenez P,
Landry C, Ruiz-Cabello F, Boon T, Garrido F. Mutations of
b2-microglobulin gene result in lack of HLA class I molecules
on melanoma cells of two patients immunized with MAGE
peptides. Tissue Antigens 1988; 52:520.
7. Cabrera T, Collado A, Fernández MA, Ferrón A, Sancho J,
Ruiz-Cabello F, Garrido F. High frequency of altered HLA
class I phenotypes in invasive colorectal carcinomas. Tissue
Antigens 1998; 52:114.
8. Lopez-Nevot MA, Cabrera T, Higuera B de la, Ruiz-Cabello
F, Garrido F. Production and characterization of monoclonal
antibodies against leukemic cells. Inmunología 1986; 5:51.
9. Neefes JL, Ploegh HL. Allele and locus-specific differences in
cell surface expression and the association of HLA class I
heavy chain with b2-microglobulin: differential effects of
inhibition of glycosylation on class I subunit association. Eur
J Immunol 1988; 18:801.
10. Orita M, Suzuki Y, Sekiya T, Hayashi K. Rapid and sensitive
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
detection of point mutations and DNA polymorphisms using
the polymerase chain reaction. Genomics 1989; 5:874.
Cordón-Cardo C, Dalbagni G, Sáez GT, Oliva MR et al. p53
mutations in human bladder cancer: genotypic versus phenotypic patterns. Int J Cancer 1994; 56:347.
Bicknell DC, Rowan A, Bodmer WF. b2-Microglobulin gene
mutations: a study of established colorectal cell lines and
fresh tumors. Proc Natl Acad Sci U S A 1994, 9:4751.
Gyapay G, Morisette J, Vignal A, Dib C, Fizames C,
Millasseau P, Marc S, Bernardi G, Lathrop M, Weissenbach J.
The 1993-1994 Genethon human genetic maps. Nat Genet
1994; 7:246.
Ramal LM, Feenstra M, Zwann AW van der, Collado A,
Lopez Nevot MA, Tilanus M, Garrido F. Criteria to define
HLA haplotype loss in human solid tumors. Tissue Antigens
2000; 55:443.
Branch P, Bicknell DC, Rowan A, Bodmer WF, Karran P.
Immune surveillance in colorectal carcinoma. Nat Genet
1995; 9:231.
Wang Z, Cao Y, Albino AP, Zeff RA, Houghton A, Ferrone S.
Lack of HLA class I antigen expression by melanoma cells
SK-MEL-33 caused by a reading frameshift in b2 microglobulin messenger RNA. J Clin Invest 1993; 91:684.
Hicklin DJ, Dellaratta DV, Kishore R, Liang B, Kageshita T,
Ferrone S. Beta2-microglobulin gene mutations in human
melanoma cells: molecular characterization and implications
for immune surveillance. Melanoma Res 1997; 7[Suppl 2]:S67.
Jimenez P, Cantón J, Concha A, Cabrera T, Fernández M, Real
LM, García A, Serrano A, Garrido F, Ruiz-Cabello F.
Microsatellite instability analysis in tumors with different
mechanisms for total loss of HLA expression. Cancer
Immunol Immunother 2000; 48:684.
Restifo NP, Esquivel F, Kawakami Y, Yewdell J, Mulé JJ,
Rosemberg SA. Identification of human cancers deficient in
antigen processing. J Exp Med 1983; 177:265.
Cromme FV, Airley J, Heemels MT, Ploegh HL, Keating PJ,
Stern PL, Meijer CJ, Walboomers JM. Loss of transporter
protein, encoded by the TAP-1 gene, is highly correlated with
loss of HLA expression in cervical carcinomas. J Exp Med
1994; 179:335.
Yano O, Kanellopoulus J, Kieran M, Le Bail O, Israel A,
Kourislky P. Purification of KBF1, a factor binding to both H2 and B2-microglobulin enhancers. EMBO J 1987; 6:3317.
Singer DS, Maguire JE. Regulation of the expression of class
I MHC genes. Crit Rev Immunol 1990; 10:235.
Blanchet O, Bourge JF, Zinszner H, Tatari Z, Degos L, Paul P.
DNA binding of regulatory factors interacting with MHC-classI gene enhancer correlates with MHC-class-I transcriptional
level in class-I-defective cell lines. Int J Cancer 1991; 6:138.
Drew PD, Lonergan M, Goldstein ME, Lampson LA, Ozato K,
Mcfarlin DE. Regulation of MHC class I and β2-microglobulin
gene expression in human neuronal cells. J Immunol 1993;
150:3300.
Feenstra M, Rozemuller E, Duran K, Stuy I, Tweel J van den,
Slootweg P, Weger R de, Tilanus M. Mutation in the b2m gene
is not a frequent event in head and neck squamous cell carcinomas with aberrant HLA class I expression. Hum Immunol
1999 60:697.