Download - Lymphoproliferative disease of 'lak cell' precursor large granular lymphocytes in association with celiac disease

American Journal of Hematology 43:116-122 (1993)

Lymphoproliferative Disease of “LAK Cell” Precursor Large Granular Lymphocytes in Association With Celiac Disease

P. Lopez, D.G. Morris, P.R. Galbraith, D.P. Lillicrap, and H.F. Pross Departments of Medicine (P.L.. P.R.G.), Microbiology and Immunology (D.G.M., H.F.P.), Oncology (P.R.G., H.F.P.), and Pathology

(D.P.L.), Queen’s University, Kingston, Ontario, Canada

We have investigated a case of lymphoproliferative disease of large granular lymphocytes (LDGL) occurring in association with celiac disease, anemia, neutropenia, and carcino- mas of the endometrium, breast, and skin. The large granular lymphocyte (LGL) proliferation was monoclonal, T cell in origin, with T cell receptor P-chain gene rearrangement, and a CD3’, CDB’, CD16’ phenotype. In spite of the high frequency of LGL, natural killer (NK) cell activity was absent. Stimulation with interleukin-2 in vitro, however, resulted in high lymphokine-activated killer (LAK) cell activity against NK-resistant targets. The T-cell nature of the LAK precursor cells is in contrast to the majority seen in normal peripheral blood. Therapeutic trials of cyclosporin A, low-dose cyclophosphamide, and levamisole were unsuccessful in reducing transfusion requirements. This case is unique in the association of LDGL with celiac disease. It is also unique in that the patient had been followed for several years prior to the onset of the LDGL. The case extends the list of lymphoproliferative disorders documented to be associated with celiac disease and, conversely, adds to our knowledge of lymphoproliferative disorder of LGL and its “dysim- mune” manifestations.

Key words: natural killer cells, interleukin-2, lymphokine-activated killer cells

(0 1993 Wiley-Liss, Inc.

INTRODUCTION

The occurrence of lymphoproliferative disorders such as non-Hodgkin’s lymphomas in patients with celiac disease has long been recognized [ I ] . Although this was initially labelled “malignant histiocytosis” [2,3], the advent of immunohistochemistry and the ability to identify T-cell receptor gene arrangements have made it clear that many of these lymphomas are of T-cell origin [4-61. The term “lymphoproliferative disease of granular lymphocytes” (LDGL) embraces a heterogeneous group of disorders characterized by atypical lymphocytosis, blood cytopenias, and, occasionally, significant organ infiltration. The disease usually exhibits an indolent course but may have a more aggressive evolution in - 15% of cases. The CD3’4-8+ 16+ (Ty) phenotype with variable coex- pression of natural killer (NK) cell markers and function is most commonly described. A number of case reports and reviews on this entity have been published recently

The heterogeneity of large granular lymphocyte (LGL) lymphoproliferative diseases with regard to cellular mor- phology. surface phenotype, and lytic capabilities makes generalizations difficult, if not impossible. The advent of

0 1993 Wiley-Liss, Inc.

[7- 161.

molecular analysis at the DNA level has allowed more precise delineation of cellular lineage and clonality . One study involving 34 LDGL patients showed that the majority had a proliferation of lymphocytes that exhibited a rearrangement within the P-chain region of the T-cell receptor (TcR) [9]. The cytotoxic potential of the lymphocytes has been observed in many studies, and several have suggested active suppression of hematopoiesis by the proliferating cells [7-9,13.15,17,18]. In the first reported case of LDGL with NK cell activity, evidence was presented that the patient’s neutropenia was caused by an antibody-dependent cell-mediated cytotoxicity (ADCC) reaction involving the proliferating N K cells and antineu- trophil serum antibody [ 191.

In this report we describe a patient with celiac disease who developed LGL lymphocytosis with very low NK

Received for publication October 14, 1991; accepted October 15, 1992.

Address reprint requests to Dr. H. Pross, Department of Microbiology and Immunology. Queen‘s University, Kingston. Ontario, Canada K7L 3N6.

Case Report: Lymphoproliferative Disorder and Celiac Disease 117

Fig. 1. Peripheral blood smear showing granular lyrnphocytosis.

activity accompanied by high lymphokine-activated killer (LAK) activity after ex vivo interleukin-2 (IL-2) incubation. This patient is of interest with respect to the LAK precursor subpopulations involved, their role in the associated dysimniune manifestations, and the association between celiac disease and LDGL.

CASE REPORT AND METHODS

A 54-year-old woman first presented in March, 1981, with stage I, poorly differentiated adenocarcinorna of the endometrium. She had no gastrointestinal symptoms at that time, and, while the complete blood count and differ- ential leukocyte counts were entirely normal, the blood smear was reported as “suggestive of postsplenectomy changes” due to the presence of Howell-Jolly bodies and target cells. In the preoperative intravenous pyelogram, the spleen could not be visualized. Family history was remarkable in that her mother had died aged 58 years of carcinoma of the uterus, and her father had died at age 57 years with multiple sclerosis. The patient was treated by total abdominal hysterectomy/oophorectomy and postop- erative radiotherapy and remained well for 5 years.

In November, 1982, she was referred to the Hernatol- ogy Department because of a low serum folate level ( 1 ng/ml; normal range 2-14 ng/ml) and a low serum B12 ( 122 pg/ml; normal range 200-1,100 pg/ml). Hematolog- ical data were normal (hemoglobin 143 @liter, WBC 8 X lO’/Iiter, neutrophils 63%, lymphocytes 30%, inonocytes 4%, eosiniphils 3%). She admitted to fatigue as her only symptom, but had experienced diarrhea, with four or five loose stools per day during the months of

September and October. The question of hyposplenism was formally addressed at that time. A liver spleen scan showed only a small amount of radioactivity in the splenic area.

She remained well until September, 1986, when she presented with fatigue. She appeared chronically ill . but there were no skin lesions, lymphadenopathy, or orga- nomegaly . Hematological data were as follows: hemoglobin 105 @liter, MCV 119 fl , leukocytes 10.3 x lo’/ liter, with 83% lymphocytes, 13% neutrophils, and 4% monocytes. Platelets were 696 X 109/liter, and reticulo- cytes were 19/1,000 erythrocytes. The blood smear showed many atypical, granular lymphocytes (Fig. I ) as well as the hyposplenic features. Marrow aspirate revealed mild normoblastic erythroid hyperplasia with a minor population showing slight megaloblastic changes. Giant band forms were noted, and there was an infiltration of mature lymphocytic cells (25%). Karyotype analysis (marrow and peripheral blood) of both unstimulated and phytohemagglutinin-stimulated lymphocytes was norinal. Serum vitamin B 12 and serum and red cell folate were low. In the Schillings test, malabsorption of vitamin B 12 was not corrected by administration of intrinsic fac- tor. Replacement therapy with vitamin B12 and folate eliminated the megaloblastic features but did not cure the anemia.

In October, 1987, she returned, giving a 7 month history of increasing diarrhea, weight loss, and fatigue. The hemoglobin was 62 giliter, and the MCV was 117 tl. A repeat marrow aspiration showed 60%) lymphocytes. There was hypoalbuniinemia and a polyclonal increase in IgA (5 .55 giliter). A small bowel biopsy revealed total

118 Case Report: Lopez et al.

Fig. 2. Jejunal biopsy showing evidence of celiac disease.

60

50 - i & 40

30

, 7

X

cn -1

0

v

20

10

0

+

+

200

175

150

_I 125 2 100

-

m 7 5 9 50

25

3 80 81 82 83 84 85 86 87 86 89 90 91

DATE

Fig. 3. Changes in hemoglobin (o), total WBC counts (+), and lymphocyte counts (A) over the period 1981-1990.

villous atrophy, and a mononuclear infiltrate, predomi- nantly of plasma cells, was present in the lamina propria, consistent with the diagnosis of celiac disease (Fig. 2). No Gitrrdiu organisms were seen in the biopsy. A CT scan showed no intraabdominal mass. She was transfused with 2 units of packed red cells and placed on a gluten- free diet. The diarrhea resolved, but the macrocytic anemia persisted despite folate and vitamin B 12 therapy. In March, 1988, she presented with a squainous cell carcinoma of the skin of the face, which was treated with radiotherapy.

In September, 1988, the anemia and lymphocytosis worsened (Fig. 3), and she became transfusion dependent. There was no palpable or ultrasonic evidence of lymphadenopathy. Bone marrow biopsy now showed ex-

tensive lymphocytic infiltration and marked erythroid and granulocytic hypoplasia. Direct Coomb’s test and HTLV-I serology were negative. HIV serology was equivocal but was negative using the polymerase chain reaction. In the fall of 1989, she was treated surgically for stage I infiltrating ductal carcinoma of the breast. Se- quential therapeutic trials using oral levamisole ( 1 50 mg, 2 consecutive days per week, January-June, 1989) and cyclosporin A (100 mgm, b.i.d., November, 1989- March, 1990) were unsuccessful in decreasing transfusion requirements. The patient’s white cell count returned to normal levels during this period, but a relative lymphocytosis (80%), neutropenia (14%), and anemia (67 g/liter) remained. In March, 1990, the patient developed cryptococcal meningitis, which was successfully treated with tluconazole. Cyclosporin A was discontinued. and low-dose cyclosphosphamide (50 mg/day) was begun, again with no effect on the hematological picture. In late April, 1990, she developed marked ascites and died on June 1 I , 1990. Autopsy showed widespread peritoneal carcinomatosis with ascites. The pattern of spread and histology were more suggestive of endometrial than of breast adenocarcinoma. The blood vessels and bone marrow showed evidence of lymphoproliferation, but no lymphocytic infiltration of the liver was reported. The brain and meninges were normal. Intestinal tissue was autolyzed, and the histological status of the celiac disease at death could not be determined.

Over the course of the patient’s disease, a number of studies were carried out to analyze the characteristics of the lymphoproliferation. As can be seen from Figure 3, the lymphoproliferative disorder was first documented in 1986, coincident with the anemia. At the onset of lymphocytosis (8.6 X IO”/liter), the cells were characterized

Case Report: Lym

550

6- 500 0 2 450

3 300

f ’ 5 0

400

2 350

250 ; 200

j 100 W 0 50

0 NK NK-IFN ADCC LAK(K5621 LAKRAJI)

TYPE OF CYTOTOXICITY

Fig. 4. NK, LAK, a-IFN-stimulated (NK-IFN) (125 U/ml) and antibody-dependent cellular cytotoxicity (ADCC) activity exhibited by the patient in February, 1988. Cytotoxicity was measured in an 18 hr 5’Cr-release assay against the two targets, K562 (NK sensitive, LAK sensitive) and RAJl (NK resistant, LAK sensitive), as previously described [20]. ADCC was measured using a rabbit anti-P815/P815 assay. The two controls were selected from a stable of 300 normal individuals because they consistently represented high and low NK activity, respectively. Lytic units were calculated as previously described [21]. Solid bars, control 1 ; cross- hatched bars, control 2; diagonally hatched bars, patient.

as CD2 94%, CD8 85%, and CD4 9%. The mean values of CD markers measured seven times between March, 1986, and death was CD3 94%, CD4 4%, CD8 89%, CD16 13%, CD25 5%. CD56 10%. The CD3+8+4- phenotype was not altered by therapy in spite of fluctua- tions in total WBC count.

In February, 1988, cell-mediated cytotoxicity was as- sessed by “Cr release, as previously described [2&-22]. Results (Fig. 4) indicated low NK activity and relatively high LAK activity after incubation with IL-2 compared with controls. Figure 4 also indicates that the stimulating effect of a-interferon on her lymphocytes in vitro was minimal, proportionally similar to that of the controls. This contrasted with the marked increase in cytotoxicity exhibited in antibody-dependent cellular cytotoxicity (ADCC) assays compared with the controls.

When the changes in NK and LAK activity were mon- itored between 1987 and 1990, a general trend towards a reduction in LAK activity was seen (Fig. 5). The peak LAK activity seen in the latter part of 1987 correlates well with the onset of the lymphocytosis. However, when fluorescence-activated cell sorting (FACS) analysis was done on the lymphocytes obtained on the dates indicated, little change was observed in the relative lymphocyte subpopulation frequency; i.e., the majority of the cells were CD3+CD8+ at all times tested. HLA determination using a microlymphocytotoxicity test showed the presence of the HLA-B8 haplotype.

Analysis of the T-cell receptor (TCR) P-gene locus was performed using standard techniques. DNA was ex-

iphoproliferative Disorder and Celiac Disease 119

3500

5 2500 - 1 3000

10 9 1500

1350

87 88 89 90 9 1

DATE

Fig. 5. PEL isolated from the patient and stored frozen were used to monitor NK and LAK cytotoxic changes between 1987 and 1990 [22]. Measurement of cytotoxicity was as described for Figure 4. The PEL collected in April, 1990, were used fresh. +, NK; A, LAK(K562); 0, LAK(Raji).

tracted from blood lymphocytes and a lymph node by treatment with sodium dodecyl sulfate-proteinase K followed by phenol-chloroform extraction as previously described [23]. DNA was digested with the restriction en- donucleases EcoRI, HindlII, and BamHI prior to agarose gel electophoresis, Southern transfer, and hybridization with a radiolabelled probe from the TCR P-gene constant region. These studies demonstrated a clonal rearrangement at this TCR gene locus (Fig. 6).

DISCUSSION

In this case, a number of features typical of lymphoproliferative disease of LGL were observed, specifi- cally, anemia, neutropenia, and malignant disease in the presence of atypical lymphocytosis of >2.0 x IO’/liter of lymphocytes with azurophilic cytoplasmic granules (normal <0.5 X 10”lliter) [ 151. Polyclonal hypergam- maglobulinemia has also been documented in association with LDGL [24]. As with the majority of 151 cases reviewed recently by Pandolfi et al. [ 151, the LGL in this case were CD3, CD8 Fcy-receptor positive, i.e., Ty cells.

The case is unusual because of the association of celiac disease with LDGL, which has not been previously reported. While we cannot exclude the possibility that both disorders were secondary to endometrial carcinoma, this seems unlikely given the natural history of the diseases. The patient was genetically predisposed to celiac disease by carrying the HLA B8 haplotype [25]. demonstrated the typical features of celiac disease, and the diarrhea resolved when she was placed on a gluten-free diet. We suspect that the disease was active though asymptomatic for many pears. The evidence for this is indirect, however. Splenic atrophy, common in celiac disease. was


Fig. 6. Autoradiograph showing the results of Southern analysis of genomic DNA extracted from the patient’s lymphocytes and a lymph node. This EcoRl digest has been hybridized with a probe from the TCR p-constant region. Two clonally rearranged hybridization bands are clearly identified, and the germline band is diminished in intensity.

present in 1981 before she was treated for stage 1 endometrial carcinoma.

Small bowel and other non-Hodgkin’s lymphomas in association with celiac sprue have recently been characterized as being of T-cell origin [4-61. The basis for the increased risk of non-Hodgkin’s lymphomas and other gastrointestinal malignancies in celiac patients is unclear. The intraepithelial lymphocytes in celiac small bowel biopsies are CD8+ but CD4-, while CD4 positivity pre- vails in the lymphocytes of the lamina propria [26]. An inverse correlation between circulating T-cell numbers and jejunal intraepithelial lymphocytes has been suggested as a theoretical basis for impaired immune surveil- lance and development of malignancy in these patients L2.5). Marsh et al. [27] have reported an increase in the absolute numbers of intraepithelial granular lymphocytes in patients with celiac disease compared with various control populations. Other studies have documented the ability of these cells to mediate NK activity [28,29], but the proportion expressing CD16 or CD57 is low [29]. Our patient’s lymphoid proliferation was phenotypically CD8+, similar to the lymphocytes found intraepithelially in sprue patients. The association of celiac disease with

LCL lymphocytosis is unusual, and the question of which came first is difficult to answer since both disease pro- cesses are frequently indolent. However, given the natural history of celiac disease, it is likely that celiac disease was present in this case for some years prior to diagnosis of endometrial carcinoma, and both the small bowel disorder and the carcinoma of the endometrium appear to have antedated the LDGL. That the patient had these two diseases prior to developing LDGL supports the conten- tion in the literature that LDGL could be reactive in nature [15]. Although the association of solid tumors with lymphocytosis of LGL is known, nonleukemic lymphoma-like presentations of LGLs have been infre- quently reported [ 17,30,31].

In 1986, the lymphocytosis was found to be of CD8+ T-cell origin and was subsequently further characterized as noted in this report. In 1988, with development of a high transfusion requirement and demonstration of marked hypoplasia of erythroid and granulocytic ele- ments in the marrow, a link was suspected between LGL and marrow failure. The anemia of LDGL is thought to be due to the inhibition of erythroid-colony-forming units and burst-forming units-erythroid by factors released from the LCL [8,13,18]. However, therapeutic interventions were without success. Neither cyclosporin A nor levamisole nor low-dose cyclophosphamide decreased transfusion requirements, in spite of a reduction in LGL count associated with the first two interventions.

The etiology of LDGL is unknown. HTLV-I and Ep- stein-Barr viruses have been linked as causative agents (7,181, although the majority of cases do not involve these agents. One recent report has suggested that the LGL proliferation is in reaction to virally infected CD4 lymphocytes (in this case, hepatitis B virus) [ 3 2 ] . Disor- ders other than celiac disease have also been associated with LDGL [IS]. These include rheumatoid arthritis [9], malignant lymphoma and fibrous histiocytoma [ 171, and chronic infections such as hepatitis B, tuberculosis, bru- cellosis. salmonellosis, and cholecytitis [9,10]. Bacterial infections are usually due to the profound neutropenia that is more commonly seen in these patients [ 151, whereas fungal infections may be due to immunosuppres- sion caused by the disease per se or to its treatment.

The lymphocytes from the patient showed CD3, CD7 (data not shown), and CD8 positivity and T-cell P-chain gene rearrangement. Compared with well-characterized controls, the patient had low NK activity. However, upon incubation with IL-2, the neoplastic cells exhibited high levels of LAK activity. In addition, high levels of ADCC activity were observed. confirming previous reports on these patients [ 18,191. The relationship between NK cells and the cells that exhibit LAK activity in the presence of IL-2 has received much attention [reviewed in 331. It is now generally agreed that LAK activity can be generated from both CD3+ and CD3- populations. In

Case Report: Lymphoproliferative Disorder and Celiac Disease 121

normal peripheral blood lymphocytes (PBL), however, the majority of LAK potential is found in the CD3-CD5 6+CD16+ population that also contains max- imal NK activity 1341. In this patient the LAK activity was derived from the non-NK CD3+ T-cell population. Unlike the NK-deficient, CD3’ LDGL reported by Bal- las et al. [35], the cells were also CD8+, CD16’. The correlation between the advent of the lymphocytosis and peak LAK generative potential in 1987 indicates that the lymphocytosis involves a LAK precursor cell population. I t is interesting that the LAK generative potential waned, subsequently, while the lymphocytosis continued. It is possible that the cells became less differentiated as the disease progressed and, concomitantly, became less re- sponsive to IL-2. There was not, however, a change in the proliferating population’s phenotype or IL-2 receptor status. In the later stages of disease, treatment with cyclosporin A and cyclophosphamide may have contributed to the decrease in IL-2 responsiveness and LAK cell cytotoxicity .

In conclusion. a case of T-cell LDGL has been described in which clinical observations preceded lymphocytosis. The disease was typical in many respects but was unusual in being associated with celiac disease and in its non-NK, LAK precursor cell activity, thus extending the spectrum of malignancies associated with celiac disease.

ACKNOWLEDGMENTS

We thank Mrs. P. Bandy-Dafoe and G. Lawrance for excellent technical assistance, Dr. Mark Minden (Ontario Cancer Institute) for the initial analysis of T-cell receptor gene rearrangements, and Miss Dianne Robertson for typing the manuscript. The work was supported by grants from the Leukemia Research Society and Physicians’ Services Incorporated (P.R.G.) and from the Cancer Re- search Society, the National Cancer Institute of Canada, and the Medical Research Council of Canada (H.F.P.).

REFERENCES

I . Fairle NH, Mackie FP: The clinical and biochemical syndrome in lymphadenoma and allied diseases involving the mesenteric lymph glands. Br Med J 1:375-380. 1937.

2. Issacson P. Wright DH: lntehtinal lymphoma associated with malabsorption. Lancet I :67-70, 1978.

3. Swinson C. Slavin G. Coles EC, Booth CC: Coeliac disease and malignancy. Lancet 1:111-115, 1983.

4. Issacson PG. Spencer JO, Connolly CE, Pollack DJ. Stein H, O’Connor NTJ, Bevnn DH. Kirkham N, Wainscoat JS. Mason DY: Malignant histiocytosia of the intestine: A T-cell lyniphorna. Lancet 2:688-69 I 1985.

5 . Ormann W. Hinchmann WD. Alexandrakis E, Bauerle H: T-cell lymphoma associated with sprue. Deutsch Me3 Wochenschr I14:13241327. 1989.

6. Seddon DJ. Chung KF. Paradina FJ, Newlands ES, Snashall PD: Long teriii survival after trcatnient of disseminated T-cell lymphoma pre-

senting with tracheal obstruction in a patient with coeliac disease. Thorax 44:5 19-520, 1989.

7. McDaniel HL. MncPherson BR, Tindle BH. Lunde JH: Lymphopro- liferalive disorder of granular lymphocytes: A heterogeneous disease. Arch Pathol Lab Med 116.242-248. 1992.

8. Oshimi K: Granular lymphocyte proliferative disorders: Report of 12 cases and review of the literature. Leukemia 2:617-627, 1988.

9. Semenzato G . Pandolfi F, Chisesi T: The lymphoproliferative disease of granular lymphocytes: A heterogenous disorder ranging from indolent to aggressive conditions. Cancel- 643297 1-2978, 1987.

10. Loughran TP. Starkehaum G: Large granular lymphocyte leukemia: Report of 38 cases and review of the literature. Medicine 66:397405, 1987.

1 1 . Chan WC, Winton EF, Waldmann TA: Lymphocytosis of large granular lymphocytes. Arch Intern Med 156:1201-1203, 1986.

12. Gastl G. Rumpold H . Kraft D, Gattringer C. Schuler G , Margreiter R, Schmalzl F. Huher C: Abnormal expansions of granular lymphocytes: Reactive lyniphocytosis or chronic leukemia? Case report and literature review. Blut 52:73-89, 1986.

13. Reynolds CW. Foon KA: T gamma-lymphoproliferdtive disease and related disorder\ in humans and experimental animals: A review of the clinical. cellular, and functional characteristics. Blood 64: 1146-1 158. 19x4.

14. Newland AC. Catousky D.Linch D. Cawley JC: Chronic T-cell lymphocytosis: A review of 21 cases. Br J Haematol 58:433446, 1984.

15. Pandolfi F, Loughran TP. Starkehaum G . el al.: Clinical course and prognosis of the lymphoproliferative disease of granular lymphocytes-A multicenter study. Cancer 65:341-348. 1990.

16. Lin CK. Liu HW. Tae PW. Lai CL. Chan GT: A patient with large granular lymphocytosis of unusual phenotype and polymorphic T-cell receptor beta-chain gene rearrangement. Am J Clin Pathol 94:211- 216, 1990.

17. Pandolfi F. Pezzutto A , DeRoss G: Characterization of two patients with lymphomas of large granular lymphocytes. Cancer 53:44.5452, 1984.

18. Oshimi K. Oshimi Y . Ahutsu M, Takei Y, Saito H. Okada M, Mi- mguchi H: Cytotoxicity of interleukin 2-activated lymphocytes for leukemia and lymphoma cells. Blood 68:938-948, 1986.

19. Pross HF, Paler J , Dwosh I , Giles A. Gallinger LA, Ruhin P, Corhett WEN, Galbraith P. Baines MG: Studies of human natural hiller cells. I l l . Neutropenia associated with unusual chiiriicteristics of antihody- dependent and natural killer cell-mediated cytotoxicity. J Clin [nimu- no1 2:12&134. 1982.

20. Morris DG, Pross HF: Studies of lyniphokine-activated killer (LAK) cells. I . Evidence using novel monoclonal antibodies that most human LAK precursor cells share a common surface marker. J Exp Med 169:717-736. 19x9.

21. Pross HF. Baines MG, Ruhin P. Shragge P. Patterson M: Spontaneous human lymphocyte-mediated cytotoxicity against tuiiiour target cells. IX. Quantitation of natural killer cell activity. J Clinical Immunol 15-63 , 1981.

22. Pross HF, Maroun JA: The standardization of NK cell assays for use in studies ot biological response modifiers. J linmunol Methods 68:235- 249. 1986.

2 3 . Maniatis T . Fritsch EF. Sambrooh 1: ”Molecular Cloning: A Labora- tory Manual.” Cold Spring Harhor. NY: Cold Spring Harhor Lahora- tory. 1982.

24. Bassan R, Pronesti M, Buzzetti M. Allavena P. Ranihaldi A. Manto- viini A. Barhui T: Autoimmunity and B-cell dydunclion in chronic proliferative disorders of large granular lympliocytes/iiatural killer cells. Cancer 63:9&95. 1989.

25. Piris J: “Coeliac Disease in Gastrointestinal and Oesophageal Pathol- ogy.” New York: Churchill Livingstone, 1989.

26. Selby WS, Janossy G . Bofill M. Jewel1 DP: Lymphocyte suhpopula- lions in the human m a l l intestine. Clin Exp Immunol52:219-228. 1983.

27. Marsh MN. Leigh RJ. Loft DE. Garner GV, Gordon DB: Studies of


intestinal lymphoid tissue X. Observations on granular epithelial Iyni- phocytes (gEL) in normal and diseased human jejunum. Virchows Arch A 412365-370. 1988.

28. Tagliabue A. Befw AD. Clark DA, Birnenstock 1: Characteristic5 of natural killer cell, in niurine intestinal epithelium and lamina propria. J ExpMed 155:1785-1796. 1982.

29. Cerf-Bensussan N. Guy-Grand D. Griscelli C: Intraepithelial lymphocytes of human gut: Isolation, characterisation and study of natural killer activity. Gut 26:81-X8. 198.5.

30. Longacre TA, Listroiii MB. Spigel JH. Willman CL, Dresslei- L: Aggressive jejunal lymphoma of large, granular lymphocytes: Imrnu- nohistochemical, ultrastructural, molecular and DNA content analysis. Am J Clin Pathol93: 124-132, 1990.

3 I , Landay A, Poon MC. Clement LT, Grossi CE: A lyniphoproliferative disorder of granular lymphocytes with novel phenotype and suppressor function. J Clin Iiiimunol 4:32&332. 1984.

32. Agohtini C, Zambello R . Pontisso P. Alberti A. Trentin L, Siviero F. Foa R . Pandolfi F, Semenzato G: Lymphoproliferative disease of granular lymphocytes in a patient with concomitant hepatitis B virus infection ofCD4 lymphocytes. J Clin Inimunol 9:401408. 1989.

33. Oilaldo J R . Longo DL: Human natural lymphocyte effector cells: Definition, analysis of activity. and clinical effectiveness. JNCl XO:999-1010. 1988.

34. Phillips JH. Lanier LL: Dissection of the lymphohine-activated killer phenomenon. Relative contribution of peripheral blood natural killer cells and T lymphocytes to cytolysis J Exp Med 1 6 4 : 8 1 ~ 8 2 5 , 1986.

35. Ballas ZK, Turner JM, Turner DA. Goetzman EA, Kemp JD: A patient with simultaneous absence o l “classical” natural killer cells (CD3 . CD16+. and NKHI+) and expansion of C D 3 ’ . CD4-. C D V . N K H l + subset. J Allergy Clin lnimunol X5:453359. 1990.