Esterase D assay in Brazilian retinoblastoma families

American Journal of Medical Genetics 34391-396 (1989)

Esterase D Assay in Brazilian Retinoblastoma Families Eugenia Costanzi, Maria E. da Silva-Fernandes, VAnia D’Almeida, and Clelia M. Erwenne Laboratory of Biochemical Genetics, Department of Morphology, Escola Paulista de Medicina (E.C., M.E. da S-F., V.D’A.)and Department of Ophthalmology, Hospital A.C. Camargo (C.M.E.), S& Paulo, SP, Brazil

The gene related to retinoblastoma (Rb gene) can be considered a model human tumor supressor gene and was assigned to band 13q14, together with the esterase D (ESD) gene. We studied the ESD activity and phenotype in 40 retinoblastoma patients, 50 unaffected relatives, and 85 nonrelated healthy control individuals. ESD activity from patients is significantly different from that of relatives and control individuals, but there was no significant difference between ESD activity from unaffected relatives and control individuals. Twelve and one-half percent of patients and 4.2% of unaffected relatives with ESDl phenotype showed a low ESD level. The results showed the importance of ESD studies in all retinoblastoma patients and their relatives. KEY WORDS: Rb gene, eye cancer, chromosome 13 marker, tumor susceptibility INTRODUCTION Damage to the genetic information of both normal alleles of the retinoblastoma (Rb) gene on chromosome band 13q14 favors the development of retinoblastoma, a malignant childhood tumor arising from retinoblasts. Other nonocular tumors, such as osteosarcoma, are also associated with inactivation of the Rb gene [Cavenee et al., 1983; Dryja et al., 19841. On the other hand, some patients may develop retinoma, which is a benign form of mutation of the Rb gene [Connolly et al., 19831. Retinoblastoma can be hereditary or sporadic. The primary difference between these forms is that hereditary cases exhibit a mutated copy or absence of the Rb gene in all cells. These patients can transmit the disease to their offspring [Cavenee et al., 19831.The small group

Address reprint requests to Dr. Maria E. da Silva-Fernandes, Department of Biochemistry, Escola Paulista de Medicina, CP 20372, SBo Paulo, CEP 04023, SP, Brazil. Received for publication January 12, 1989; Revision received April 10, 1989.

0 1989 Alan R. Liss, Inc.

of retinoblastoma patients who lack the Rb gene usually shows only 50%of esterase D (ESD)activity, because the ESD gene is also located on the 13q14 band and is tightly linked to the Rb gene [Sparkes et al., 1980,19831. ESD is an enzyme found in many human tissues, but its biological role is unknown. This esterase is a polymorphic enzyme with three phenotypes: ESDl, ESD2-1, and ESD2, resulting from the expression of two alleles: ESDl and ESD2 [Hopkinsonet al., 19731. The activity is measured by hydrolysis of 4-methylumbelliferyl acetate (MUA), and the level of the ESDl phenotype is higher than that of the ESD2-1 or ESD2 phenotype [Nishigaki et al., 1983; Horai and Matsunaga, 19841. The quantitative ESD assay has been used to identify the integrity of the 13q14 band in individuals with retinoblastoma and their relatives [Junien et al., 1982;Turleau and de Grouchy, 19871and to screen populations of patients with retinoblastoma associated with loss of the Rb gene [Cowell et al., 1986al. This assay has also been used to detect deletion carriers not previously diagnosed by cytogenetic techniques [Benedict et al., 1983; Cowell et al., 19871. Since the segregation of ESD alleles corresponds to the inheritance of the Rb gene, the ESD phenotype can be used as a marker of the presence of a mutated Rb gene and can help predict individuals at high risk of developing tumors in families with hereditary retinoblastoma [Sparkes et al., 1983; Mukai et al., 1984;Halloran et al., 19851. Recently,apolymorphic site for Apa I, located within the ESD gene, was used to identify the chromosome 13 carrying the defective Rb allele in families with retinoblastoma [Young et al., 19881. In this study we examined the ESD phenotype and activity in a series of retinoblastoma patients, and unrelated healthy individuals from Brazil.

SUBJECTS AND METHODS Selection of Samples The 40 retinoblastoma patients consisted of 18 (45%) sporadic cases and 22 (55%)familial cases. There were 25 (62.5%)with unilateral tumor and 15 (37.5%)with bilateral tumors. The patients were seen either for follow-up or as new cases at the Ophthalmologic Department, A.C. Camargo Hospital, Siio Paulo, Brazil. Tumor diagnosis in-

392

Costanzi et al.

cluded complete ophthalmologic and histopathologic examination. Twenty-one of the patients were accompanied by relatives and 19 were not. Fifteen families were examined; 11families with hereditary retinoblastoma and 4 families with the sporadic form. In five of the families with hereditary cases, relatives were affected: F1 to F5 (Fig. 1).On the other hand, relatives with retinoma and osteosarcoma were observed in two pedigrees (F6 and F7, respectively). Thus, the presence of these phenotypes indicated a mutation in the Rb gene. Families F8 to F11 contained bilateral retinoblastoma cases. Most unaffected relatives (35 of familial cases and 15 relatives of sporadic cases) also underwent eye examination. The control group consisted of 85 unrelated healthy individuals without a family history of retinoblastoma.

E

:fi2-lh ESDl 165

ESDl 45

ESDZ-1

i

ESDl

98

09

The age range was 6 months to 30 years for patients, 9 to 61 years for unaffected relatives, and 7 months to 53 years for normal control individuals.

Sample Processing Peripheral blood samples for ESD testing were drawn in heparin from 175 individuals and processed immediately. Erythrocytes were washed in saline, packed, and lyzed by freezing and thawing. Esterase D Analysis ESD was initially typed by conventional electrophoresis using 12% starch gel (or 1%agarose) in 0.1 M sodium phosphate, pH 6.5, at 20 Vlcm for 4 hours at 4°C. The gel buffer was 1 : l O of the electrode. Samples were also analyzed by low-voltage agarose

X

l

ESDl 61

ESD%I ESDl 43 31

e

ESD2-1 114

F7

113

79

113

F9 ESD1

bL ESDl

ea

ESDl 107.5

ffli tE;A ESD 1 70

I801 141

ESOl 14s

ESDl

IS01

I3

83

ESDl

88

*&YlL*

ESD1 151

E

i

P

ESD1

ESDI

01

112

OD

Y

6-a

ESDl 88

ESDl

ESDl

102

ESDl 181

ESDl

€801

13.

ps

ESDl 179

Fig. 1.Pedigrees of retinoblastoma families demonstrating ESD activity and phenotype, 0 0, unaffected retinoblastoma;[l@, unilateral retinoblastoma (right eye); 0 0, unilateral retinoblastoma (left eye);m, bilateral retinoma; U, osteosarcoma. 0, bilateral

Esterase D in Retinoblastoma Families isoelectric focusing (AGIEF). AGIEF was carried out using gel prepared as 1%agarose-IEF (Pharmacia, Sweden), 12% sorbitol, and 7% ampholyte, pH 4.0-6.5 (Pharmacia). Agarose and sorbitol were first dissolved in twice-distilled water and cooled to 75"C, and Pharmalyte was added just prior to gel casting. Erythrocyte lysates ( 2 ~ 1were ) applied to the gel at a distance of 2 cm from the cathode. The anode had 0.01 M HzS04 and the cathode 1 M NaOH. The distance between electrode wicks was 13 cm. Focusing was carried out for 45 min a t constant power (15W) and then addtionally for 45 min a t constant voltage (1,000 V and unlimited mA). The cooling temperature was 4°C. ESD patterns were demonstrated according to Hopkinson et al. [19731. Erythrocyte ESD activity was measured by MUA hydrolysis using the fluorimetric method described by Sparkes et al. [1980].A separate experiment with repeat samples did not show variation. As a routine, samples were analyzed in duplicate within a 1 month period. A normal control group was included in each series of experiments. The levels are reported as mol MUA hydrolyzedlhrlg of hemoglobin. Statistical Analysis Samples were divided into three main groups: retinoblastoma patients, unaffected relatives, and normal control individuals. Each group was subdivided into other small groups according to ESD phenotype, cause, and laterality of the tumor. ESD activity did not follow a gaussian distribution. Thus, results were expressed as the natural logarithm of each ESD reading to improve the symmetry of the distributions. The geometric means and standard deviations were calculated for all groups and subgroups. Means were compared by the Mann-Whitney test. RESULTS The ESDl phenotype was the most frequent in all groups, and no difference in allele frequency was found among the three main groups. No case of a rare ESD allele was identified. Segregation of the Rb gene was followed by ESD phenotype in two hereditary retinoblastoma families: F2 and F6 (Fig. 1). The results of ESD activity and phenotype are shown in Table I. There were significant differences among the ESD activities for the phenotypes in the normal control group. Individuals with the ESDB allele consistently showed low ESD activity. However, only groups and subgroups with the same phenotype were compared. Significant differences were also found between ESDl of retinoblastoma patients and ESDl of normal control individuals (P = 0.0128) and between ESDl of retinoblastoma patients and ESDl of unaffected relatives (P=0.0319). No significant differences were seen between: 1)unaffected relatives and normal control individuals: 2) unilateral and bilateral cases: and 3) sporadic and hereditary cases. The comparison of ESD2-1subgroups did not show a significant difference possibly because of the small number of cases. The reference value for detecting the cases with low ESD activity was the geometric mean of the normal

393

TABLE I. Summed Scores for ESD Activity and Phenotype of Groups and Subgroups Phenotype ESDl ESD2-1

Geometric meana 86.4 74.1

Unaffected relatives

ESDl ESD2-1

Control

ESDl ESDB-1 ESDB

Group Patients

"Units x hemoglobin.

1.6 1.6

No. 33 7

105.7 78.7

1.5 1.2

47 3

108.6 76.9 50.0

1.4 1.4 1.9

72 10 3

SD

mol of methylumbelliferyl acetate hydrolizedihrlg of

control group for each phenotype. The geometric mean of ESDl activity in the normal control group was 108.6 units and was taken as 100%ESD activity. All individuals with levels below the 54.3 units (50%of the mean) were considered as carriers of only one copy of ESD and of the Rb gene. The same criterion was used to define ESDB-1phenotype. For these individuals activity levels below 38.5 units were considered deficient. Figure 2 shows the distribution of values expressed as a percentage of the control group's ESD mean for each phenotype. According t o this criterion, we identified four retinoblastoma patients and two unaffected relatives with low ESD levels in their blood cells (Table 11). No case was detected in the group of healthy individuals.

DISCUSSION Here we report the results of the first quantitative and qualitative ESD analysis carried out on a group of Brazilian patients with retinoblastoma and their relatives. Segregation of the Rb gene was followed in two families with hereditary cases. In family F2, the mother had a bilateral retinoblastoma and ESDB-1 phenotype, and her daughter (11-2) had a unilateral tumor and ESDl type. Thus, the mutated Rb gene was co-segregatedwith the ESDl allele. In family F6, the father (bilateral retinoma) and his child (111-1) with unilateral retinoblastoma both showed the ESDB-1 phenotype. Thus, the ESDB allele accompanied the mutated Rb gene [Costanzi et al., 19891. Since the pioneering work by Sparkes et al. [19801, ESD analysis has been used as an important marker of the integrity of chromosome band 13q14. Individuals with low ESD activity can be identified by comparing their ESD levels with those of relatives or with the mean values detected in a normal control group. Although a family study of ESD represents the most adequate and informative procedure, it is not always possible to perform enzyme analyses in the entire family. Thus, the selection of a normal reference value for ESD activity becomes of fundamental importance. The ESDl phenotype exhibits higher enzyme activity than ESDB-1 or ESDB, especially among control individuals. This observation emphasizes the importance of phenotypicESD determination in parallel with fluorimetric analysis. The patient group studied in this report showed ESD activity significantly different from that

394

i

Costanzi et al. 71

RETINOILASTOMA

PATIENTS

1

L

1 1 1

UNAFFECTED RELATIVES

v) YI

v)

<

U IL

0 a

1

1 CONTROL

m 0 6 ? 6 5 4

3 2 1

4

% OF NORMAL MEAN

Fig. 2. Distribution of ESD levels in 175 individuals from Brazil. The arrows indicate the limit value. Means of control group: ESD1, 108.6; ESDB-1, 76.9; ESD2. 50.0.

detected in relatives and control individuals. Thus, we selected 50% of the normal mean value for each phenotype a s the minimum level of the ESD activity. This criterion permitted the identification of four patients with low ESD activity, a result compatible with hemizygosis of genes Rb and ESD.These were two cases of hereditary retinoblastoma and two cases with the sporadic form of the tumor. After ESD analysis, the latter two cases were also considered to be hereditary. Among the four patients with low ESD activity, only the RB2 case (Fl,11-21 had a positive family history. This patient has bilateral retinoblastoma and only 41.1% of the normal level of ESD activity and inherited the tumor from his father (Fl,1-11, who had normal enzyme activity. All tested relatives had the ESDl phenotype. The RB2’s father may have the following ge-

notype in his cells: ESD1, ESD1,Rb/ - , - , i.e., both alleles are in chromosome 13 with the wild Rb allele. This could be a plausible explanation for the above findings. Thus, RB2 inherited chromosome 13 with deficiency of the Rb and ESD loci. Among the relatives, we detected two individuals with reduced ESD activity, i.e., F4,111-2 and F10,1-2, sister and mother of the patient with hereditary retinoblastoma, respectively. Three hypotheses may be proposed in a n attempt to explain these results.

1.Firstly, these individuals might have a deficiency in 13q14 that involved both the ESD and Rb genes but did not undergo the second mutation of Rb gene. According to the KnudsonlCavenee model, this would trigger the tumor, or alternatively, the tumor cells failed to pro-

Esterase D in Retinoblastoma Families TABLE 11. Cases With Reduced ESD Level”

Case RB2 RB19 RB27 RB28 UR8 UR31

Retinoblastoma ESD assay % of normal Pedigree type position Etiology Laterality Unitsb mean 45.0 41.1 F1,II-2 H U H B 52.0 47.9 S U 44.0 40.5 S U 20.0 18.4 F4, 111-2 No No 31.0 28.5 37.5 34.5 NO No F1O.I-2

“RB,retinoblastoma patient; UR, unaffected relative; H, hereditary; S, sporadic; U, unilateral; B, bilateral. bunits x mol of 4-methylumbelliferyl acetate hydrolizedihrlg of hemoglobin.

gress due to some unknown cause. In fact, the absence of retinoblastoma in individuals with low ESD activity has been reported previously [Pankau et al., 1987;Wilson et al., 1987; Cowell et al., 19881. 2. Another tenable hypothesis is that these relatives might have a null allele or ESDO. Recently, Lee et al. 119871 discussed the presence of the ESDO allele in a patient with retinoblastoma and low ESD activity. The mutation that gives origin to the ESDO allele may be independent or linked to the cause of retinoblastoma through an unknown mechanism. 3. Finally, it is also possible that these relatives may be hemizygous only for ESD gene. The fact that individuals with low ESD activity were found among relatives of hereditary retinoblastoma cases indicates the presence of susceptibility to mutation in the Rb locus. We cannot state that 4.2% of the relatives with the ESDl phenotype are hemizygous for the Rb gene. However, the presence of low ESD levels in relatives of hereditary cases points to the existence of a mutation close to the Rb gene. This mutation may be transmitted to the children of these individuals together with susceptibility to retinoblastoma. We observed a small number of individuals with high ESD activity in the three main groups studied. A similar result was reported by Cowell et al. [1986bl, suggesting that duplication of the ESD gene may be a natural process. These cases may be more sensitive to mutation in the 13q14 band. Few studies have been conducted using serial ESD measurements in patients with retinoblastoma. One of the first studies was described by Dryja et al. 119831, who analyzed 51 patients. More recently, Cowell et al. l1986al measured ESD activity in 200 patients, 9 of whom showed decreased ESD levels. These authors did not consider phenotypic differences and defined cases with ESD levels of less than 70%the mean value (205.3 units) as 13ql4-deficient. On the basis of this criterion, most of our patients and their relatives, as well as normal individuals, could be considered deficient. In contrast, Cowell et al. [1986b], in a study of ESD activity of 330 individuals (258 patients with retinoblastoma and 72 relatives), obtained a mean of 103.7 and 101units for the ESDl and ESD2-1 phenotypes, respectively. These values are similar to those obtained here, but Cowell et al. [1986b] assigned all individuals to a single group

395

maintaining the cutoff of 70% of the mean as the reference value. Also combining relatives with healthy control individuals in the reference group, van der Heiden et al. [19881 considered that the ESD screening of all patients was open to question. The present results suggest that each laboratory intending to use ESD measurement for analysis of the number of copies of band 13q14 should carefully establish its own criteria and reference values. Each country should have a reference center considering the genetic components of its population. We also think that low or high ESD activity or presence of the ESDO allele could be considered as “scars” left by the process of activation or inactivation of one or more genes in band 13q14, including the Rb gene, during normal cellular development. Thus, individuals with these characteristics may belong to a subgroup with greater susceptibility to mutation in Rb gene and therefore to the development of retinoblastoma and osteosarcoma and perhaps other tumors, such as of the mammary duct or of small lung cells IMarx, 19881. Thus, the identification of individuals with mutation of the gene is becoming increasingly important. At present, molecular analysis of the Rb gene can identify most of these individuals, as reported by Friend et al. [19861 and Wiggs et al. [19881. However, such a sophisticated technology is not yet applied routinely and thus, judicious ESD analysis may be an important tool for the identification of individuals with Rb gene deficiency. The latter technique can be applied routinely to all patients and their relatives in modestly equipped laboratories and at a reduced cost. It could also be used to help screen those cases that deserve to be analyzed by molecular techniques.

ACKNOWLEDGMENTS The authors are indebted to Dr. E.L.F. Franco, Ludwig Institute for Cancer Research, Si50Paulo, for the statistical analysis. We also thank Dr. J.A.D. Andrade, Department of Morphology, Escola Paulista de Medicina, Si5o Paulo. This study was supported by grants from FAPESP, CNPq and CAPES, Brazil. REFERENCES Benedict WF, Murphree AL, Banerjee A, Spina CA, Sparkes CA, Sparkes RS (1983): Patient with 13 chromosome deletion: evidence that the retinoblastoma gene is a recessive cancer gene. Science 219:973-975. Cavenee WK, Dryja TP, Phillips RA, Benedict WF, Godbout R, Gallie BL. Murahree AL. Strong LC. White RL (1983): Exmession of recessive‘alleles by chrom&omal mechanisms in retinoblastoma. Nature 305:779-784. Connolly MJ, Payne RH, Johnson G, Gallie BL, Allderdice PW, Marshall WH, Lawton RD (1983):Familia1,EsD-linked, retinoblastoma with reduced penetrance and variable expressivity. Hum Genet 65:122-124. Costanzi E, Silva-Fernandes ME, Erwenne, CM (1989): Esterase D analysis in familial retinoma and retinoblastoma. Ophthalmic Paediatr Genet, in press. Cowell JK, Rutland P, Jay M, Hungerford J (1986a): Deletions of the esterase D locus from a survey of 200 retinoblastoma patients. Hum Genet 72:164-167. Cowell J K , Rutland P, Jay M, Hungerford J (198613): Effect of the esterase-D phenotype on its in vitro enzyme activity. Hum Genet 74:298-301,

396

Costanzi et al.

Cowell JK, Thompson E, Rutland P (1987): The need to screen all retinoblastoma patients for esterase D activity: detection of submicroscopic chromosome deletions. Arch Dis Child 62%-11. Cowell JK, Rutland P, Hungerford J, Jay M (1988):Deletion of chromosome region 13q14 is transmissible and does not always predispose to retinoblastoma. Hum Genet 80:43-45. Dryja TP, Bruns GAP, Gallie B, Petersen R, Green W, Rapaport JM, Albert DM, Gerald PS (1983): Low incidence of deletion of the esterase D locus in retinoblastoma patients. Hurn Genet 64:151-155. Dryja TP, Cavenee W, White R, Rapaport JM, Petersen R, Albert DM, Bruns GAP (1984): Homozygosity of chromosome 13 in retinoblastoma. N Engl J Med 301:550-553. Friend SH, Bernards R, Rogelj S, Weinberg RA, Rapaport JM, Albert DM, DryjaTP (1986):A human DNA segment with properties ofthe gene that predisposes to retinoblastoma and osteosarcoma. Nature 323:643-646. Halloran SL, Boughman JA, Dryja TP, Mukai S, Long D, Roberts DF, Craft AW (1985):Accuracy of detection of the retinoblastoma gene by esterase D linkage. Arch Ophthalmol 103:1329-1331. Hopkinson DA, Mestriner MA, Corter J , Harris H (1973) Esterase D a new human polymorphism. Ann Hum Genet 37:119-137. Horai S, Matsunaga E (1984):Differential enzyme activities in human esterase D phenotypes. Hum Genet 66:168-170. Junien C, Despoisse S, Turleau C, Nicolas H, Picard F, le Marec B, Kaplan J-C, de Grouchy J (1982):Retinoblastoma, deletion 13q14, and esterase D: application of gene dosage effect to prenatal diagnosis. Cancer Genet Cytogenet 6281-287. Lee W-H, Bookstein R, Wheatley W, Benedict WF, Lee EY-HP (1987): A null allele of esterase D is a marker for genetic events in retinoblastoma formation. Hum Genet 76:33-36. Marx J L (1988):Eye cancer gene linked to new malignancies. Science 241:293-294. Mukai S, Rapaport JM, Shields JA, Augsburger JJ, Dryja TP (1984):

Linkage of genes for human esterase D and hereditary retinoblastoma. Am J Ophthalmo 97:681-685. Nishigaki I, Itoh T, Ogasawara N (1983): Quantitative variations in polymorphic types of human red cell esterase D. Ann Hum Genet 47:187-192. Pankau R, Johannson W, Grote W, Doner K, Tolksdorf M (1987): Interstitial deletion of chromosome 13 involving the region 13q14. Hum Genet 77:292-293. Sparkes RS, Sparkes MC, Wilson MG, Towner JW, Benedict W, Murphree AL, Yunis JJ (1980):Regional assignment of genes for human esterase D and retinoblastoma to chromosome band 13q14. Science 208:1042-1044, Sparkes RS, Murphree AL, Lingua RW, Sparkes MC, Field LL, F’underburk SJ, Benedict WF (1983): Gene for hereditary retinoblastoma assigned to human chromosome 13 by linkage to esterase D. Science 219:971-973. Turleau C, de Grouchy J (1987): Constitutional karyotypes in retinoblastoma. Ophthalmic Paediatr Genet 8:ll-17. van der Heiden C, Geurtzen AFM, Brink W, de Klerk JBC, Beemer FA (1988):Esterase D: evolution of a potential derived gene marker for hereditary retinoblastoma. Clin Chim Acta 17323-88. Wiggs J, Nordenskjold M, Yandell D, Rapaport J , Grondin V, Janson M, Werelius B, Petersen R, Craft A, Riedel K, Liberfarb R, Walton D, Wilson W, Dryja TP (1988): Prediction of the risk of hereditary retinoblastoma, using DNA polymorphisms within the retinoblastoma gene. N Engl J Med 318:151-157. Wilson WG, Campochiaro PA, Conway BP, Carter BT, Sudduth KW, Watson BA, Sparkes RS (1987): Deletion (13) (q14.lq14.3) in two generations: variability of ocular manifestation and definition of the phenotype. Am J Med Genet 28:675-687. Young L-JS, Lee EY-HP, To H, Bookstein R, Shew J-Y, Donoso LA, Sery T, Giblin M, Shields JA, Lee W-H (1988):Human esterase D gene: complete cDNA sequence, genomic structure, and application in the genetic diagnosis of human retinoblastoma. Hum Genet 79:137-141.