BritishJourrtal of Hautnatology, 1981, 47, 21 1-220.

HLA Antigens and Childhood Acute Lymphocytic Leukaemia

FREDERICK R. DAVEY, NEIL A. LACHANT, NANCY L. DOCK, CHARLENE HUBBELL, JAMES A. STOCKMAN, I11 A N D JOHN B. HENRY

Departments oj‘Pathology, Medicine, and Pediatrics, S . U. N. Y. Upstate Medical Cenrer, Syracuse, New York

(Received 15 April 1980; acceptedforpublication 10Jirne 1980)

SUMMARY. HLA-A and -B antigens were determined for 94 children with acute lymphocytic leukaemia (ALL) and for 376 normal controls. Sixty-four of these 94 patients were typed for lymphocyte surface markers and 59 were defined as ‘null cell’ ALL. There was no difference in the distribution of the HLA-A or -B locus antigens between the control group and the entire group of patients with ALL or the ‘null cell’ subgroup. Patients with HLA-A9 determinants had a significant increase in early, first remission duration compared to patients without HLA-A9. This was particularly evident in the ‘null cell’ ALL subgroup. In addition, HLA-A9 appeared to be an independent factor affecting the length of first remission since there was no correlation between known prognostic factors such as patient age, sex or WBC and the presence or absence of the HLA-A9 antigen. Survival for the first 12-18 months was also greater in the HLA-A9 group than in the non-HLA-A9 population. Thus, the presence of HLA-A9 appears to be associated with some protective effect among patients with ALL.

A relationship was demonstrated in mice between the composition of genes a t the major histocompatibility locus and resistance to some forms of virus-induced leukaemia (Lilly, 1966, 1968, 1972). As a result, several studies were conducted in humans to determine the relationship between the HLA system and acute lymphocytic leukaemia (ALL). No consistent observations have been made. In some studies there appeared to be a correlation between certain HLA antigens and the occurrence of acute lymphocytic leukaemia (Thorsby et al , 1971; Walford et a l , 1970; Rogentine et al , 1973; Sanderson et a l , 1973; Gluckman et al , 1976; Hester et a l , 1977; Albert et al , 1977). However, in other reports no association was observed (Lawler et a l , 1971, 1974; Batchelor er al, 1971; Dick et al , 1972; Klouda et al , 1974; Terasaki & Mickey, 1975; Cohen et a l , 1977). In addition, the presence of HLA-A2 (Rogentine et al , 1973; Dausset, 1977) and HLA-A9 (Lawler et al , 1974; Cohen et al , 1977) has been reported to be associated with an increase in long-term survival in patients treated with chemotherapy. Unfortunately many of the previous studies were performed retrospectively (Harris & Lawler, 1978). Thus, patients

Correspondence: Dr Frederick R. Davey, Department ofPathology, S.U.N.Y. UpstateMedical Center, 750 East Adams Street, Syracuse, New York 13210, U.S.A.

0007-1048/81/0200-0211$02.00 0 1981 Blackwell Scientific Publications

211

212

who died early may not have been HLA typed, and a distortion of the results could have occurred.

Three major subtypes of ALL are recognized on the basis of lymphocyte surface markers. Patients with T cell ALL represent approximately 20% of cases and have circulating blasts which form nonimmune rosettes with sheep erythrocytes (Brouet et ul, 1976; Sen ei Borella, 1975). B cell ALL forms a second group and occurs in 1-5% of cases (Davey 8; Gottlieb, 1974). I n this subtypc of ALL, lymphoblasts have surface-membrane-bound immunoglobulin. A third group ha5 been referred to as ‘non-T, non-B’ cell or ‘null cell’ ALL and has been observed in a t least 75% ofpatimts (Brouet et a / , 1976; Davey & Gottlieb, 1974). Lymphoblasts from this group do not form E rosettes nor contain surface-membrane-bound immunoglobulin. As yet, no studies have examined the relationship between these subdivisions of ALL and the distribution of HLA determinants.

The current study describes the results at the Upstate Medical Center of a group ofpediatric patients with ALL who were typed prospectively for HLA antigens and lymphocyte surface markers, treatcd according to standardized protocols (Cancer and Leukemia Group B), and observed for LIP to H years.

Frederick R. Davey et a1

MATERIALS A N D METHODS

P u r i c m uiid Cmtrok Thc study group consisted of 94 white children, less than 18 years of age, treatcd a t the

S.U.N.Y. Upstate Medical Center between 1971 and 1979. A diagnosis of ALL was based on the presence of lymphoblasts in the peripheral blood and bone marrow. The blasts had no peroxidase (Kaplow, 1975) or Sudan black B positive (Sheehan 8: Storey, 1947) cytoplasmic granules. However, in most cases a variable number of blasts contained coarse periodic acid-Schiff (PAS) positive granules (Hayhoe 8: Flenians, 1969).

All typing for lymphocyte surface markers was done at the time of initial diagnosis. In 66 of 94 (70%) cases. HLA typing was performed on peripheral blood lymphocytes collected during the initial month of first remission. In 28 of 94 (30%) cases, the HLA typing was accomplished a t diagnosis. HLA genotypes were performed on 46 of 94 (49%) patients providing verifica- tion of the typing results. If extra reactions were observed during initial phenotyping, then the family was studied and the genotype was determined on the patient. The criteria for diagnosis, remission, and relapse of ALL were those of Cancer and Leukemia Group B (CALGB, 1974). Seventy-one patients were treated according to CALGB protocols (Nos. 71 11, 741 1, 761 1).

A group of 376 normal healthy subjects unrelated to the study group served as controls. HLA-A and -B lymphocyte antigens were determined, using antisera obtained from the

NIH (National Institutes of Health) employing a standard NIH microlymphocytotoxicity test (Ray et a / , 1976).

Lyaipliocyte Su!fuce M a r k e r s In 64 patients, mononuclear cell suspensions were tested for nonimmune rosette formation

with sheep erythrocytes (E) according to the methods of Bach (1973) and Jondal et nl (1972). Surface-membrane-bound immunoglobulin w a s detected according to the technique of Papamichail et a/ (1971).

H L A Antigerrs and Childhood A L L 213

Statistical Analysis The frequencies of HLA phenotypes from patients with ALL were compared to the control

population utilizing the chi-square function. Life table calculations and their statistical analysis were performed by standard methods (Colton, 1974). Cumulative differences in life table calculations were analysed by the logrank method (Petro et al, 1977).

RESULTS

Patients with A L L There were 44 males and 50 females. The median age a t diagnosis was 6.4 years and all

patients were less than 18 years of age. Ninety-two of 94 children had a complete remission after induction chemotherapy.

The frequency of HLA phenotypes in 94 children with ALL did not differ significantly (P> 0.05) from those of the control population (Table I). Similarly, there was no difference in the distribution of HLA phenotypes of 44 patients who survived for more than 36 months when compared to those of 26 individuals who died within the first 36 months after the

TABLE I. Frequencies of HLA antigens (percentages)

All patients Patients with with ALL 'tiul/' ce/I ALL Controls

Antigen (N=94) (N=59) (N=376)

A1 A2 A3 A9

A10 A1 1 A19 A28 A29 A30 A3 1 A32

B5 B7 B8

B12 B13 B14 B15 B16 B17 B18 B27 B35 B40

28.7 48.9 26.6 234 11.7 9.6

15.0 4.3 3.2 1 .o 0 0

21.3 30.9 17.0 25.5 4.3 4.3

11.7 3.2

11.7 9.6 4.3 6.4

14.9

23.7 50.8 22.0 27.1 13.6 15.2 15.2 5.1 3.4 1.7 0 0

20.3 32.3 10.2 27.1 3.3 3.3

10.2 3.3

15.2 5.1 5.1

10.2 20.3

33.0 51.6 26.1 19.1 13.0 7.7

15.2 2.1 2.1 0.5 0.3 0

13.8 22.3 23.9 21.5 5.1 4.5

14.6 2.9 7.2

10.4 9.6 4.5

12.0

214 Frederick R. Davey et a1

TAME 11. Frequency of H L A antigens (perccnrages)

Sirruival less Survival more HLA tllarr 36 rttoriths tliatt 36 tnotitks

arrtigrrl (S= 26) (,Y=44)

A1 A2 A3 A9

A10 A1 1 A19 A28 A 29 A30 A3 1 A32

B5 B7 B8

B12 B13 B14 B15 B16 B17 B18 €327 €333 B40

3 4 6 17.3 34.6 19.7 7.7 7.7

23.1 0 0 0 0 0

34.6 308 23.1 34.6

3.8 7.7 7 7 0

11.5 11.5 0 0

1 5 4

-- 33.7 50.0 29.5 -- 37.7 13.6 6.8

15.9 6.8 0 0 0 0

-- 77.7 29.5 18.2 25.0

4.5 2.3

15.9 2.3 9.1 6.8 4.5 7 .3

13.6

diagnosis of ALL (Table 11). Twenty-four patients who are still alive, but who have not been observed for a t least 36 months, were omitted from this comparison.

Since other studies had noted prolonged survival in patients with either HLA-A2 or HLA-A9 determinants, we analysed the duration of first remission and the length of survival following a complete response to chemotherapy after stratification of the study population for the presence or absence ofHLA-A2 and A9 antigens. No significant difference was observed in the actuarial duration of first remission or length of survival when patients with HLA-A2 were compared with individuals without HLA-A2 (Fig 1). Conversely, the presence of HLA-A9 appeared to be of prognostic importance (Fig 2). A significantly higher percentage of patients with HLA-A9 were in a continuous complete response at 6 months (100% I > . 91%. P<O41) and a t 12 months (100% 1’. 77%, P<0.003) after remission induction when compared to those individuals who lacked the HLA-A9 antigen. After 18 months of continuous remission, however, the advantage for those who had the HLA-A9 antigen was lost. There was no significant difference in the median duration of remission between the two groups, nor was there a significant difference in overall remission duration by logrank analysis. Although the percentage of survivors was greater for the group with HLA-A9, this was statistically

HLA Antigeris and Childhood ALL

DURATION OF SURVIVAL

-

- - - - - -

215

:2/N"ll' I! 244. 22

I 6 $;

lo&.. II

"L ' 3 2 I 6 8 5 4 4 ----- L

3 1

I I l I I , I I

L Q, n

All Patients with A L L

a_--- H LA- A 2

20 HLA-A2 "1 -Non

I I I I I I I I I I

Pat ients with 'Null' Cell A L L

4 4 3 1

Patients with 'Null Cell' ALL Lymphoblasts from 64 of 94 patients were evaluated for the presence of lymphocyte surface

markers. Lymphoblasts from 59 of the tested patients failed to form E rosettes o r to have surface-bound immunoglobulin and these cases were classified as 'null cell' ALL. The fre- quency of HLA-A and -B determinants from these 59 cases of 'null cell' ALL are recorded in

226 Frederick R. D a v e y et a1

DURATION OF SURVIVAL

+ c W 0 L 0, a

Al l Patients with A L L

80 - 70 -

:s 4 60 - .__. 50 -

40-

30 -

\.7 7 7 .. . _. . . .

6 5 H L A - A 9 .____. - Non H L A - A 9

Patients with

L.... 2 1

1I DURATION O F FIRST REMISSION

A l l Patients with A L L

W

a 2 0 $ 10

Patients with ‘Null’ Cell A L L

i! .... 1 1

10 20 30 40 50 60 70 80 90 100 10 20 30 4 0 50 60 70 80 Months Months

FIG 2. Duration of survival and first remission for patients with ALL. Patients with HLA-A9 had a significantly better duration survival and first remission than patients with non-HLA-A9.

Table I. In patients with ‘null cell’ ALL, HLA-B8 was observed to be less frequent than controls (P<0.02), whereas HLA-B17 was noted to be more frequent in patients with ALL than in the control population (P< 0.04). However, when these probabilities were multiplied by the number of HLA specificities, the significance of these differences no longer existed.

There was no significant difference in the duration of first remission or subsequent survival of the patients with ‘null cell’ ALL. when stratified for the presence or absence ofHLA-A2 (Fig 1) . Again, the presence of HLA-A9 appeared of prognostic importance (Fig 2). Significantly more patients with HLA-A9 were in a continuous response at 6, 12 and 18 months after remission induction when compared to those who lacked the HLA-A9 antigen. For those individuals who remained in remission for more than 18 months the selective advantage for the

H L A Atitlgens arid Childhood A L L

TABLE 111. Comparison of other prognostic factors in ‘null’ cell ALL according to the presence or absence of HLA-AY

217

~~

HLA-A9 N O ~ I HLA-A9

Sex Males Females Total

Age Mean +_ SD

< 2 years 2-8 years <8 years

WBC Mean SD Median

< 20 x 1 0y/1 2cklOo x 109/1 > 1oox 10”l

Unknown

~~

Number

P>0.05 5 16

Years 6.3 +_ 4.6

Number

;} P> 0.05 5

x I @ / / 21.1+25.8 P>0.05

5.2

Numher

P> 0.05

2

Number

I ;: 43

Years 6 . 8 k 4.4

Number 5 I :2”

{ 2:

x 10911

60.2& 105 10.6

Number

7

0

presence of HLA-A9 disappeared, since there was no significant difference in overall remission duration by logrank testing at 66 months.

Although the actuarial survivorship was higher for those with HLA-A9, survival was significantly increased only at 18 months. At 66 months there was no difference in overall survival by logrank analysis.

To determine if the prolongation of first remission observed in patients with HLA-A9 was the result of some other prognostic factor, patients with ‘null cell’ ALL were also compared according to sex, age, and WBC (Table 111). No significant differences were observed in the distribution of these parameters between patients with and without the HLA-A9 antigen.

Patierits with T and B Ce l l A L L

small number did not allow for a meaningful statistical analysis. Five patients had lymphoblasts with lymphocyte markers characteristic o f T or B cells. This

DISCUSSION

Kourilsky et al (1967) was the first to study the relationship between HLA antigens and ALL. They compared 116 patients with ALL to 234 controls and found no significant differences in the distribution of the HLA determinants between these two populations. However, Walford et al (1970) and Thorsby et a! (1971) noted that the haplotype HLA-A2 B12 was increased among patients with ALL. The current study and numerous other studies (Batchelor et al,

218 Frederick R . Davey et a1

1971; Harris R Viza, 1971;Jeannet & Magrin, 1971; Davey etal , 1974; Klouda et a/, 1974) failed to demonstrate a significant relationship between the occurrence of ALL and the presence of HLA-A2 or HLA-B12. In addition, the current study showed no association between HLA-A2 determinants and patients with ‘null cell’ ALL.

Rogentine er a / (1973) observed that patients with HLA-A2 and HLA-A2, HLA-A9 were over-represented in individuals who survived longer than 1500 d. Therefore they suggested that an increased frequency of HLA-A2 in patients with ALL may be the result of a prolonged survival rather than an increased susceptibility to leukaemia. However, Gluckman et al(1976) and Albert et nl(1977) observed an increased frequency of HLA-A2 in patients with ALL at the time of presentation. In addition, the current study demonstrated no significant increase in the duration of first remission or in survival for all patients with ALL or for individuals with ‘null cell’ ALL Lvho had the HLA-A2 determinants. Thus, the reason for these discrepancies is not yet clear. Perhaps there exists only a weak association between HLA-A2 and ALL and this relationship is only observed in very large studies. However, environmental risk factors may interact with individuals bearing the HLA-A2 antigen and thus alter the frequency of these antigens among patients with childhood ALL in one but not another locality. In addition, other factors may change the length of survival of patients with a particular HLA haplotype. DeBruyere el a/ (1980) observed prolonged survival in ALL patients with HLA-A2 B12 and HLA-A2 B4O haplotypes when all patients were treated with transfer factor.

LaLvler et a1 (1974) reviewed a group of 58 children with ALL and noted that those with HLA-A9 had the longest median survival. Cohen ct al (1977) observed 21 patients with ALL. He noted that seven of 11 individuals who lived longer than the median survival had HLA-A9 antigens, whereas only two of 10 patients who lived less than the median survival had HLA-A9 determinants. In addition. Sanderson rt nl(1973) and Klouda et nl(l974) observed a decreased frequency of HLA-A9 antigen among patients with ALL. Thus, the hypothesis emerged that the HLA-A9 antigen was associated with resistance to ALL.

In the current study there was no significant difference in the frequency of HLA-A9 antigens between the control group and the entire group of patients with ALL or the subgroup with ‘null cell’ ALL. However, the presence of HLA-A9 appeared to be of prognostic significance for first remission duration, especlally in the subgroup of patients with ‘null cell’ ALL. Significantly more patients with the HLA-A9 antigen were in a continuous first remission at 18 months than \vere patients without HLA-A9. If, however, patients without HLA-A9 remained in a continuous remission for 18 months, their long-term prognosis was no different from that of patients with HLA-A9. The presence or absence of HLA-A9 had a less pro- nounced effect in survival, perhaps due to the heterogeneity of therapy after initial relapse.

Pdtients Lvith and without HLA-A9 ivere further stratified according to sex, age and initial leucocyte count. None of these prognostic factors appeared to be unevenly distributed. Although more patients with non-HLA-AY had higher leucocyte counts a t the time of diagnosis, statistical analysig revealed no significant difference in the distribution of higher or lo\ver leucocytc count$ in the t\vo populations. Therefore, these data support the previous studies of Lawler cr n / (1974) and suggest the presence of HLA-A9 exerts some independent protective ettect in patients Lvith ALL.

Lilly (1972) has demonstrated a linkage in mice between the H-2K histocompatibility genes and Rgv- 1 genes which code for resistance against virus-induced leukaemia. In humans, there

H L A Antigens and Childhood ALL 21 9

is little information regarding the genetic mechanisms which protect against the development of neoplastic disorders. However, it is possible that a linkage disequilibrium exists between HLA-A9 and an immune response gene providing some degree of host-resistance to patients with ALL.

REFERENCES

ALHERT, E.D., NIsPERos, B. & THOMAS, E.D. (1977) HLA antigens and haplotypes in acute leukemia. Leukemia Resrarcli, 1,261-269.

BACH, J.F. (1973) Evaluation of T cells and thymic serum factors in man using the rosette technique. Transplanfation Reuieuis, 16, 19G217.

BATCHELOR, J.R., EDWARIIS, J.H. & STUART, J. (1971) Histocompatibility and acute lymphoblastic leuk- aemia. Lancet, i, 699.

BROUET, J.C., VALENSI, F., DANIEL, M.T., FLANDRIN, G., PREUD’HOMME, J.L. & SELIGMANN, M. (1976) Immunological classification of acute lymphoblas- tic leukemias: evaluation of its clinical significance in a hundred patients. British Journal ofHaematology, 33,319-328.

CANCER A N D ACUTE LEUKEMIA B (CALGB) (1974) Criteria for evaluating acute leukemia. Junc 1974.

COHEN, E., SINGAL, D.P., KHURANA, U., GREGORY, S.G., Cox, C., SINKS, L., HENDERSON, E., FTTZPA- T R I C K , J.E. & HICHY, D. (1977) HLA-A9 and sur- vival in acute lymphocytic leukemia and myelocy- tic leukemia. H L A and Ma/ignancy (ed. by G. P. Murphy, E. Cohen, J. E. Fitzpatrick and D. Press- man), p. 65. Progress in Clinical and Biological Research 16. Alan R. Liss, New York.

COLTON, T. (1974) Statistics in Medicine. Little, Brown and Company, Boston.

DAIISSET, J. (1977) HLA and association with malig- nancy. A critical view. H L A arrd Malignancy (ed. by G. P. Murphy, E. Cohen, J . E. Fitzpatrick and D. Pressman), p. 131. Progress in Clinical and Biologi- cal Research 16. Alan R. Liss, New York.

DAVEY. F.R., HENRY, J.B. Br GOTTLIEB, A.J. (1974) HL-A antigens and acute lymphocytic leukemia. Americari Joii m a I of Cliri ica I Pathology, 61, 662-665.

DAVEY, F.R. &- GOTTLIEB, A.J. (1974) Lymphocyte surface markers in acute lymphocytic leukemia. American Journal cf Clinical Pathology, 62,8 18-822.

DEBRUYERE, M., CIIRNU, G., HEREMANS-BRACKE, T., MALCHAIRE,J. & SOKAL, G. (1980) HLA haplotypes and long survival in childhood acute lymphoblastic leukemia treated with transfer factor. British Journal of Haematology, 44, 243-251.

DICK, F.R., FORTUNY, I., THEOLOGIDES, A., GREALLY, J.. WOOD, N. & YUNIS, E.J. (1972) HL-A and lym- phoid tumors. Cancer Research, 32, 2608-261 1.

GLUCKMAN, E., LEMARCHANEL, F., NUNEZ-ROLDAN, A., HORS, J. & DAUSSET, J. (1976) Possible excess of

acute lymphoblastic leukemia. H L A arrd Distvse Predisposition to Disease and Clinical Implications (ed. by Inserm, Paris), Vol. 58, p. 226.

HARRIS, R. & VizA, D. (1971) HL-A, leukaemia, and leukaemia associated antigens. Lancet, i, 1134-1 135.

H A R R I S , R. & LAWLER, S.D. (1978) The HLA system in acute leukemia and Hodgkin’s disease. British Medi- cal Bulletin, 34,301-304.

HAYHOE. F.G.J. & FLEMANS, R.J. (1969) An Atlas (!f Haematology Cyto logy , p. 316. WolfMedical Books, London.

HESTER, J.P., ROSSEN, R.D.. TEMPLETON, J.W.. MCCREDIE, K.B. & FREIREICH, E.J. (1977) Histo- compatibility antigens in adult acute leukemia. H L A and Malignancy (ed. by G. P. Murphy, E. Cohen, J. E. Fitzpatrick and D. Pressman), p. 131. Progress in Clinical and Biological Research 16. Alan R. Liss, New York.

JEANNET, M. & MACNIN, C. (1971) HL-A antigens in malignant diseases. Transplantation Proceedirys, 3,

JONDAL, M., HOHN, G. & WIGzELL, H. (1972) Surface markers on human T and B lymphocytes. I. Large population of lymphocytes forming non-immune rosettes with sheep red blood cells.Journa1 CfExperi- mental Medicine, 136,207-215.

KAPLOW, L.S. (1975) Substitute for benzidine in mye- loperoxidase stains. American Journal of Clinical Pathology, 63, 451.

KLOUDA, P.T., LAWLER, S.D., TILL, M.M. & HARDIsrY. P.M. (1974) Acute lymphoblastic leuke- mia and HL-A. A prospective study. Tissue Antigetis, 4, 262-265.

KOURILSKY, F.M., DAUSSFT, J., FEINGOLI), N., DUPUY, J.M. & BERNARD, J. (1967) Etude de la repartition des antigens leucocytaires chez des malades atteints de leucemie aigue en remission. Aduanres in Trans- plantation (ed. by J. Dausset, J. Hamburger and G. Mathe), p. 515.

LAWLER, S.D., KLOUDA, P.T., HARDISTY, R.M. &- TILL, M.M. (1971) The HL-A system in lympho- blastic leukemia. A study ofpatients and their fami- lies. British Journal of Haematology, 21, 595405.

LAWLER, S.D., KLOUDA, P.T., SMITH, P.G., TILL, M.M. 8; HARDISTY, R.M. (1974) Survival and the HL-A system in acute lymphoblastic leukaemia. British Medical Journal, i, 547-548.

LILLY, F. (1966) The inheritance ofsusceptibility to the

1 30 1 - 1 303.

220 Frederick R. Davey et a1

Gross leukemia virus in mice. Genetics, 53,52%538. LILLY. F. (1968) The effects of histocompatibility-2

type on response to the Friend leukemia virus in mice. Jourrral qf Experimerrtal Medicirie, 127, 465-473.

LILLY, F. (1972) Mouse leukemia: a model o f a mul- tiple-gene disease. Journal q j rhe A‘utiorral Cntricr lristitute, 49, 927-934.

PAPAMICHAIL, M., BROWN, J.C. 8; HOLBOROW, EJ. (1971) Immunoglobulins on the surface of human lymphocyter. Lancet, i i , 85G852.

PETRO, R., PIKE, M.C., ARMKAGE, P., BRESLOW N.E., Cox,D.R., HOWARD, S.V.,Marvr~i, N.,McPHER- so&, K., P n ~ o , J. &. SMITH, P.G. (1977) Design and analysis of randomized clinical trials requiring pro- longed observation of each patient. 11. Analysis and examples. Bririslijoitrrial qf Cancer, 35, 1-30.

RAY, J.G., HARE. D.B., PEDERSON. P.D. & M U L L ~ L L Y . D.I. (eds) (1976) h’lAID Mariiral c j Tissirc Typiti,q Techniqircs. p. 22. National Inctitute5 of Health. Bethesda.

ROGFNTINE. G.N.. TRAPANI, R.J., YANKEE. R.A. R- HENDERTON. E.S. (1973) HL-A antigens and acutc lymphocytic leukemia: the nature of the HL-A? association. Tissire Antipis, 3, 47&476.

SANDERSON, A.R., HOSSEIN MAHOUR, G., JAEFE, N. &‘ DAS, L. (1973) Incidence of HL-A antigens in acute lymphocytic leukemia. Trarrsplantarion, 16, 672474.

SEN, L. & BORELLA, L. (1975) Clinical importance of lymphoblasts with T markers in childhood acute lymphocytic leukemia. New Englarid Jourrial of Medicitre, 292, 828-831.

SHEEHAN, H.L. & STOREY, G.W. (1947) An improved method of staining leukocyte granules with Sudan black B. Joirrrral q/ Pathology atid Bacreriology. 59, 33G337.

TERASAKI, P.I. & MICKEY, M.R. (1975) HL-A haplo- types of 32 diseases. Transp/aritation Reoiews . 22, 105-1 19.

THORSBY. E., ENCESET, A. & LIE, S.O. (1971) HLA antigens and susceptibility to disease. A study of patients with acute lymphoblastic leukemia, Hodg- kin’s Disease, and childhood asthma. Tissue Aritigeris, 1, 1477152.

W A L ~ ~ R I I , R.L., FINKELSTEIN, S., NEERHOUT, R.. KON- R A I L P. 8r SHANBROM, E. (1970) Acute childhood leukacmia in relation to the HL-A human trans- plantation genes. Narure, 225, 461-462.