Pediatric Hematology and Oncology
ISSN: 0888-0018 (Print) 1521-0669 (Online) Journal homepage: http://www.tandfonline.com/loi/ipho20
An Immunohistochemical Study of Testicular Biopsies in Childhood Acute Lymphoblastic Leukemia: Reactivity of Normal Testicular Components and Leukemic Infiltrates B. S. Wilkins, J.H Williams, J.A Kohler & D.B Jones To cite this article: B. S. Wilkins, J.H Williams, J.A Kohler & D.B Jones (1992) An Immunohistochemical Study of Testicular Biopsies in Childhood Acute Lymphoblastic Leukemia: Reactivity of Normal Testicular Components and Leukemic Infiltrates, Pediatric Hematology and Oncology, 9:4, 297-307, DOI: 10.3109/08880019209016601 To link to this article: http://dx.doi.org/10.3109/08880019209016601
Published online: 09 Jul 2009.
Submit your article to this journal
Article views: 2
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ipho20 Download by: [137.189.171.235]
Date: 22 March 2016, At: 15:44
AN IMMUNOHISTOCHEMICAL STUDY OF TESTICULAR BIOPSIES IN CHILDHOOD ACUTE LYMPHOBLASTIC LEUKEMIA: REACTIVITY OF NORMAL TESTICULAR COMPONENTS AND LEUKEMIC INFILTRATES
B. S. Wilkins, MRCPath, and J.H. Williams, FlMLS 0 University Department of Pathology, Level E, South Block, Southampton General Hospital, Tremona Road, Southampton, SO9 4XY United Kingdom
J.A. Kohler, MRCP
Downloaded by [] at 15:44 22 March 2016
0 University Department of Child Health, Level E, South Block, Southampton General Hospital, Tremona Road, Southampton, SO9 4XY United Kingdom
D. B. Jones, PhD 0 University Department of Pathology, Level E, South Block, Southampton General Hospital, Tremona Road, Southampton, SO9 4XY United Kingdom
We performed an immunohistochemical analysis of frozen sections from testicular biopsies from 23 children with acute lymphoblastic leukemia. Eleven cases were infiltrated by leukemia. Tumor cells were immunostained by a panel o j antibodies that identified CDI 0, CD43, CDI 9, CD3, CD7, and MHC class I and II. The immunoreactivily of normal testicular components was also studied. Normal testis showed no CDlO reactivi& Wide variation in the number of stromal macrophages identified by CDllc was found. Fansjerrin receptor (CD71) was expressed by some stromal macrophages, by seminifnus tubules, and by Leydig cells. B lymphocytes were absent from the testicular stroma but small numbers of T lymphocytes were consistently present. MHC class I and II were expressed by most stromal cells but not by seminiferous tubules. KEY WORDS: acute lpphobkutic leukemia, immunostaining, testis.
INTRODUCTION The testis is a frequent site of involvement by acute lymphoblastic leukemia (ALL) in childhood, with a reported incidence of relapse at this site, during or after initial chemotherapy, of 8-16%.''3 These reports have also suggested that as a result of longer survival of children with acute leukemia, the frequency of testicular relapse is increasing.
Pediatric Hmtology and Oncology, 9:297-307, 1992 Copyright 0 1992 by Hnnisphne Publishing Corporation
297
Downloaded by [] at 15:44 22 March 2016
298
6. S. WlLKlNS ET AL
Testicular involvement may be an isolated finding,"3 but often heralds systemic relapse and thus warrants aggressive therapy even if evidence of systemic disease is absent . I Consequently, bilateral testicular biopsies have been performed as part of the assessment of male children after completing 2 years of maintenance chemotherapy. The aim has been to detect testicular involvement by leukemia at a subclinical stage, in the belief that early treatment would lead to improved survival. The need for routine testicular biopsies has been questioned in recent years, but the issue remains contr~versial.~-' In general, simple light microscopic analysis of biopsy morphology has been used to assess the presence or absence of leukemic infiltration within testes. The significance of diffuse or focal minimal infiltration has not been specifically addressed. It has been suggested that immunohistochemistry may be a useful diagnostic adjunct to routine histochemical staining when morphological assessment of the testicular biopsies is equivocal. Few studies have reported immunophenotypic data for leukemic testicular infiltrates1'"' and there has not been a systematic analysis of the immunoreactivities of normal testicular components, particularly of stromal cells that could cause diagnostic confusion with leukemic cells. We have reviewed immunohistochemical staining of testicular biopsies from children with ALL in order to characterize the antigen expression of leukemic infiltrates and normal testicular cells. We have also considered the value of immunostaining for detection of minimal involvement.
PATIENTS, MATERIALS, AND METHODS From 1985 to 1990, 29 boys with acute leukemia underwent testicular biopsy at Southampton General Hospital, either for symptomatic testicular enlargement (7 cases) or at the end of 2 years of maintenance chemotherapy (22 cases). Their ages at diagnosis ranged from 7 months to 14 years (mean age 6 years and 4 months). All of the children were treated according to Medical Research Council (MRC) protocols in use at the time of their illness. Fresh tissue was available from 24 of the cases, including 11 of 12 with infiltration by ALL. Immunostaining was performed using a well-established indirect immunoperoxidase technique" on frozen sections of each sample at the time of biopsy. All cases were studied using a diagnostic panel of monoclonal antibodies directed against a range of lineage-specific/lineage-related antigens. The antibody panel varied over the 5-year period; the antigens studied are described in the Results section, classified according to the proceedings 13 of the Fourth International Leukocyte Typing Conference. The immunophenotypes of the leukemic infiltrates were reviewed and the reactivities of normal testicular components with the antisera determined. The
TESTICULAR IMMUNOSTAININC IN ALL
299
subsequent clinical progress of patients in whom there was testicular involvement was also reviewed.
RESULTS
Downloaded by [] at 15:44 22 March 2016
Prevalence of Testicular Involvement by Leukemia Testicular involvement by leukemia was found in all 7 of the symptomatic and in 6 of the routine cases. One of the patients with testicular enlargement had a diagnosis of acute monoblastic leukemia and is not included hereafter. The other 12 boys had diagnoses of ALL (8 common ALL [CALL],2 T-ALL [T cell ALL], 2 ALL not otherwise specified); fresh tissue for immunostaining was available from 11 of these 12 boys. The twelfth child had cALL with subclinical testicular involvement at the time of routine biopsy. Table 1 summarizes the clinical course of each of the 11 ALL patients for whom immunostaining was performed and shows the results of clinical, histological, and immunohistochemical evaluation of testicular involvement. In two patients occult testicular involvement was found by immunostainTABLE
1. Diagnosis and Clinical Outcome of Patients in This Study
Testicular Involvement Diagnosis
C
H
I
Clinical Coursea
1
cALL
+
2
cALL cALL
4
T-ALL T-ALL ALL cALL
+ + + + + + + + + +
?T relapse at 2 y (no Rx). BM relapse at 3 y and T at 6 y. Further T relapse at 9 y followed by BM-Death. BM, T, and CNS relapse at 2 y. Autografted. Further BM
3
+ + + + + + + + -
Case
5 6 7
8 9 10
cALL CALL cALL
11
ALL
Abbreviations: C
+ + + + +
-
-
- -
+
relapse- Death. BM relapse at 4 y, T at 5 y, and further BM relapse at 6 y Death. T relapse at 1 y, BM relapse at 2 y-Death T relapse at 2 y. BM and CNS relapse at 3 y-Death. T relapses at 3 y and 4 y. Currently in remission at 9 y. BM and T relapse at 2 y. Further BM relapse at 2 y 6 mo- Death. BM and T relapse at 2 y. Allografted. Well at 5 y. T and CNS relapse at 2 y. BM relapse at 3 y-Death. Minimal T involvement at 2 y. Well at 2 y 8 mo. Autografted at 8 y but died. CNS relapse at 3 y. Minimal T involvement at 6 y. BM relapse at 6 y 6 mo. Autografted at 8 y but died.
-
-
-
-
detected clinically (testicular enlargement); H detected histologically in routine detected by immunohistochemical staining; T testicular; BM bone
H+E stained sections; I
marrow. 'Dates given as from the time of diagnosis.
300
B. S. WlLKlNS ET AL
ing alone. One of these patients (case 11) developed bone marrow relapse within 6 months of testicular biopsy; the other (case 10) remains well 8 months from the end of his chemotherapy, One further patient (case 3) had immunohistochemical evidence of leukemia in both testes despite clinical symptoms affecting one testis only.
Downloaded by [] at 15:44 22 March 2016
Distribution of ALL Infiltrates Within the Testis Patterns of leukemic infiltration within testicular stroma are shown in Figures 1-3. Infiltration was diffuse in most case (Fig. l ) , with sparing of tubules, but two cases showed marked leukemic infiltration of tubular basement membranes (cases 2 and 4; Fig. 2). Case 1 showed a tendency for localization of the infiltrate to perivascular areas (Fig. 3) in each of two episodes of unequivocal testicular relapse separated by an interval of 3 years. lmmunohistochemicalReactivities
of ALL Infiltrates Table 2 summarizes the tumor cell immunophenotypes of the positive ALL cases. All cases expressed CD10. Individual cases of CALL were also
FIGURE 1. Diffuse interstitial infiltration of testicular stroma by leukemic cells. Only occasional leukemic cells encroach on tubules. Immunostained for CD10.
Downloaded by [] at 15:44 22 March 2016
TESTICULAR IMMUNOSTAINING IN ALL
301
FIGURE 2. Diffuse interstitial infiltration by leukemic cells with extension into the layers of the basement membrane of seminiferous tubules. Immunostained for CDIO.
FIGURE 3. Perivascular condensation of leukemic cells within testicular stroma. Immunostained for CD10.
Downloaded by [] at 15:44 22 March 2016
302
B. S. WlLKlNS ET AL
variously positive for CD22, CD19, CD43, and MHC class I and 11. No expression of immunoglobulin light or heavy chains was seen. The two TALL cases reacted only weakly with CDlO but showed strong expression of CD3 and CD43. Immunostaining for terminal deoxynucleotidyltransferase (TdT) was not generally performed on our cases because the antibody produced only a weak reaction when positive and showed widespread nonspecific staining of all testicular components. Cell proliferation within the tumor infiltrates, as indicated by reactivity with Ki67, differed considerably between cases and did not appear to correlate with the clinical course. Case 1 suffered two episodes of symptomatic testicular relapse late after initial chemotherapy and biopsies were immunostained on both occasions. Between relapses further intensive chemotherapy was given. In the first biopsy 25% of cells were reactive with Ki67, while in the second biopsy this figure was reduced to 1%. In case 11 a microscopic focus of leukemic cells, not apparent in H&E stained sections, was detected by its reactivity for CDlO (Fig. 4) and in case 10 occasional CDlO reactive cells were scattered throughout the stroma. No other positive immunohistochemical marking was obtained in these cases. The occult infiltrate in the clinically normal testis in case 3 had the same immunophenotype as that of the leukemic cells within the abnormal one. In nine cases detailed immunophenotypic data for leukemic cells from the TABLE 2. Immunophenotypes of Leukemic Infiltrates Within Testicular Biopsies Antigen Diagnosis
CDlO
CD3
CD5
CD7
CD19
CD22
CD37
CD43
MHC-I
MHC-I1 ~~
CALL T-ALL ALL nos
717 212 212
015 212 112
014 212 nt
013 212
415
011
nt
111
416 nt nt
216 011
nt
215 212 nt
414 nt nt
214 nt 011
Results are expressed as number of cases positivelnumber of cases tested. No expression of immunoglobulin light or heavy chains was found. Cells in cycle, identified using the monoclonal antibody Ki67, varied from 1 % to 75%. Expression of CD2, 4, 8, 35, 38, and 71 was also studied in a proportion of cases. Both T-ALL cases expressed CD2 and case 5 was also positive for CD4, 8, and 71 (case 4 could not be tested for these antigens due to the small biopsy size). None of the other cases tested showed expression of these antigens. Antibodies used: 55, RFAL 1, 2, and 3 (CD10); OKT3, UCHTl (CD3); OKTl (CD5); HB2 (CD7); HD37 (CD19); To15 SCHLl, HD6 (CD22); WR17 (CD37); WRl4 (CD43); HB95 (MHC I, American Type Culture Collection); 185 (MHC 11-ICRF, Loncoln’s Inn Fields, London); Ki67 Gerdes. Forschungsinstitiit, Borstel, Germany); O K T l l (CD2); OKT4, 73F11 (CD4); OKT8 (CD8); El 1 (CD35); AT13.5 (CD38); HB21 (CD71). Abbreviations: C D cluster of differentiation. CD reactivities of individual antibodies are categorized according to the proceedings of the Fourth Leukocyte Typing Conference (see Ref. 13). MHC I and I1 major histocompatibility antigens, classes I and 11. nt not tested. nos not otherwise specified.
u.
-
-
-
-
Downloaded by [] at 15:44 22 March 2016
TESTICULAR IMMUNOSTAINING IN ALL
303
FIGURE 4. Microscopic focus of CD10-positive cells in the testicular stroma of case 11.
patients' blood or bone marrow at diagnosis were available for comparison with the results of our testicular immunostaining. Although there was only partial overlap between the panels of test antisera, no inconsistencies of antigen expression were found between leukemic cells at diagnosis and at testicular relapse. lmmunohistochemical Reactivities of Normal Testicular Components
Immunostaining of uninvolved testicular biopsies showed that Sertoli cells, gonocytes, stromal macrophages, and Leydig cells express CD7 1 (Fig. 5). Sertoli cells and endothelium react with antibody directed against CD38, an antigen expressed by late B lymphocytes and plasma cells (Fig. 6). One previous report has suggested that germ cells express CD38" but in our study they were negative. Most stromal cells express MHC class I and 11. Strong M H C class I expression by endothelial cells was also seen. The number of stromal macrophages, as identified by expression of CD1 lc, varied widely between biopsies. No B lymphocytes were identified in the normal stroma, but small numbers of T lymphocytes were consistently present. CD4- and CD8-positive T lymphocytes were present in equal numbers in most cases.
Downloaded by [] at 15:44 22 March 2016
304
6. S. WlLKlNS ET A 1
FIGURE 5. Stromal Leydig cells and cells within the seminiferous tubules (Sertoli cells and developing gonocytes) expressing CD71. Tubular staining is relatively weak, but comparison with the unstained seminiferous tubules of Figures 3 and 4 enables the cytoplasmic “blush” of CD71 to be appreciated.
DISCUSSION Testicular involvement by childhood ALL is a common accompaniment of the disease but in our series it was clinically apparent in only 6 of the 12 cases. In the other 6 cases leukemic infiltration was detected only by routine microscopy of hematoxylin-eosin (H&E)-stained sections and/or immunohistochemistry, including 2 patients in whom leukemic cells were detected only by immunostaining for CD10. One further patient who had clinical symptoms of leukemia involving only one testis showed immunohistochemical evidence of additional, sparse, leukemic infiltration of the other. The number of cases is too small for statistical analysis but there did not appear to be any correlation between the histological pattern or the extent of infiltration and the subtype of ALL present. The clinical significance of the presence of small number of CD10-positive cells within testicular stroma, as in cases 10 and 11, is unclear. CDlO is not uniquely expressed by ALL blasts, having also been demonstrated on some normal hematopoietic stem cells, renal and gut epithelium, neutrophils, melanoma and glioma cell lines, and cultured fibroblast^.""^ We saw no CDlO expression by normal testicular cells in 12 definitely uninvolved cases. Therefore it is unlikely that scattered CD10-positive cells in the testis simply repre-
Downloaded by [] at 15:44 22 March 2016
TESTICULAR IMMUNOSTAINING IN ALL
305
sent a normal stromal population and it seems reasonable to assume that they are leukemic cells. It remains a possibility that following intensive chemotherapy nonneoplastic, CD 10-positive, immature lymphoid cells could be present in tissues in response to inflammatory stimuli. Against this, we saw small numbers of phenotypically mature T cells in all of the biopsies that were not positive for leukemia, in the absence of any CD10-positive cells. Such difficulties in the interpretation of minimal or equivocal infiltration raise questions about the importance of detecting subclinical testicular relapse. Since the recommendation of routine testicular biopsies in the late 1970s, 3-9 many studies have attempted to assess the value of the procedure. Several have shown that clinically normal children with histological evidence of testicular relapse have a worse prognosis than those with normal histology, but few data are available to compare the outcome of clinically overt testicular relapse with relapse detected by histology alone. It has been suggested that it may be unnecessary to detect subclinical testicular relapse since current therapy permits successful reinduction of remission once the disease has subsequently become overt .6 Other workers, however, demonstrated a survival advantage associated with detection of testicular involvement at a subclinical rather than an overt stage3 or found that detection of occult testicular relapse identifies a group of patients who have a high risk of subsequent systemic relapse.' The
FIGURE 6. Immature seminiferous tubules showing unstained gonocytes (arrowed) among CD38positive Sertoli cells. The positive staining can be appreciated best by comparison with unstained tubules in Figures 3 and 4.
Downloaded by [] at 15:44 22 March 2016
306
B. S. WlLKlNS ET A1
value of performing testicular biopsies in asymptomatic boys at the end of maintenance chemotherapy remains an issue of controversy. The problem of interpreting the occasional finding of minimal or equivocal infiltration has been noted in several of the studies mentioned above, but the outcome for patients with such findings is not known. Clinical outcome for the CALL cases in our study was variable and the numbers were too small to indicate whether specific immunophenotypes had good or bad prognostic significance. The two T-ALL cases had rapid progression of their disease, consistent with their having very high white blood cell counts at diagnosis (data not shown). In both cases testicular relapse was clinically obvious before completion of maintenance therapy. Variation in Ki67 reactivity may be an intrinsic feature of the disease in each patient and/or an effect of therapy. That the latter is contributory is supported by case 1, in whom there was a substantial reduction in cycling (Ki67-positive) cells between the two biopsies from different episodes of testicular relapse. As no normal testicular components were found to express CD10, we consider that this antigen is the most reliable of currently detectable markers for ALL infiltrates within the testis. Other antigens that might identify leukemic cells are expressed by normal stroma (eg, various T-lymphocyte- and macrophage-associated antigens, MHC class I and II), tubules, and Leydig cells (CD71). Our finding of CD38 expression by Sertoli cells and not by gonocytes would appear to be at variance with results obtained by Verdi et al," who reported that germ cells were CD38-positive. The photomicrograph with which they illustrate CD38 immunostaining strongly suggests that Sertoli cells were positive, although they make no reference to Sertoli cell reactivity in their text. The seminiferous tubules are not shown at sufficiently high magnification to identify germ cells with uncertainty, and the discrepancy may lie in the terminology used to describe tubular cells rather than in a true difference in reactivity between our two studies. B lymphocytes were not found in uninvolved testicular stroma, in keeping with our previous findings in the adult testis." As some B-cell-associated antigens (eg, CDl9 and CD22) were commonly expressed by ALL cells, these are useful additional indicators of leukemic infiltration.
REFERENCES 1. Stoffel TJ, Nesbit ME, Seymour HL. Extramedullary involvement of the testes in childhood leukemia.
Cnncn. 1975;35:1203-1211. 2 . Hustu HO, Aur RJA. Extramedullary leukaemia. Clin Hasmafol. 1978;7:313-337. 3. Bowman WP, Aur RJA, Hustu HO, Rivera G. Isolated testicular relapse in acute lymphocytic leukemia of childhood: categories and influence on survival. J Clin Oncol. 1984;2:924-929.
Downloaded by [] at 15:44 22 March 2016
TESTICULAR IMMUNOSTAINING IN ALL
307
4. Askin FB, Land VJ, Sullivan MP, et al. Occult testicular leukemia: testicular biopsy at three years continuous complete remission of childhood leukemia: a Southwest Oncology Group study. Cuncn. 1981;47:470-475. 5. Nachman J, Palmer NF, Sather HN et al. Open-wedge testicular biopsy in childhood acute lymphoblastic leukemia after two years maintenance therapy: diagnostic accuracy and influence on outcomea report from Children’s Cancer Study Group. Blood. 1990:75:1051-1055. 6. Hudson MM, Frankel LS, Mullins J, Swanson DA. Diagnostic value of surgical testicular biopsy after therapy for acute lymphocytic leukemia. J Pediatr. 1985:107:50-53. 7. Wong KY, Ballard ET, Strayer FH, et al. Clinical and occult testicular leukemia in long-term survivors of acute lymphoblastic leukemia. J Pediutr. 1980;96:569-574. 8. Ortega JJ, Javier G, Toran N. Testicular infiltrates in children with acute lymphoblastic leukemia: a prospective study. Med Pediatr Oncol. 1984;12:386-393. 9. Ise T, Kishi K,Imashuku S, et al. Testicular histology and function following long-term chemotherapy of acute leukemia in children and outcome of the patients who received testicular biopsy. Am J Pediutr H ~ m ~ t On~ol. ol 1986;8:288-293. 10. Verdi CJ, Hutter J, Grogan T M . Immunophenotyping to detect and characterize acute lymphocytic leukemia in testicular biopsies. Pediutr Padol. 1989;9: 117-130. 11. Wilkins BS, Williamson JMS, O’Brien CJ. Morphological and immunohistochemical study of testicular lymphoma. Histoputhology. 1989;15:147-156. 12. Jones DB, Moore K, Wright DH. Heterogeneity of B-cell antigen expression in non-Hodgkin’s lymphoma. Adv Exp Med Biol. 1989;137:139-144. 13. Knapp W, et al. (eds). Leukocyte Typing, Vol. 4. Oxford: Oxford University Press;1989. 14. Hoffman-Fezer G, Knapp W, Thierfelder S. Anatomical distribution of CALL antigen expressing cells in normal lymphatic tissue and in lymphomas. Leukemia Res. 1982;6:761-767. 15. Dorken B, Moller P, Pezzutto A, et al. B cell antigens: CDIO. In: Knapp W, et al, eds. Leukocyte Typing, Vol. 4. Oxford: Oxford University Press;1989:33-34.
Received October 14, 1991 AcceptedJunuury 22, 1992
ISSN: 0888-0018 (Print) 1521-0669 (Online) Journal homepage: http://www.tandfonline.com/loi/ipho20
An Immunohistochemical Study of Testicular Biopsies in Childhood Acute Lymphoblastic Leukemia: Reactivity of Normal Testicular Components and Leukemic Infiltrates B. S. Wilkins, J.H Williams, J.A Kohler & D.B Jones To cite this article: B. S. Wilkins, J.H Williams, J.A Kohler & D.B Jones (1992) An Immunohistochemical Study of Testicular Biopsies in Childhood Acute Lymphoblastic Leukemia: Reactivity of Normal Testicular Components and Leukemic Infiltrates, Pediatric Hematology and Oncology, 9:4, 297-307, DOI: 10.3109/08880019209016601 To link to this article: http://dx.doi.org/10.3109/08880019209016601
Published online: 09 Jul 2009.
Submit your article to this journal
Article views: 2
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ipho20 Download by: [137.189.171.235]
Date: 22 March 2016, At: 15:44
AN IMMUNOHISTOCHEMICAL STUDY OF TESTICULAR BIOPSIES IN CHILDHOOD ACUTE LYMPHOBLASTIC LEUKEMIA: REACTIVITY OF NORMAL TESTICULAR COMPONENTS AND LEUKEMIC INFILTRATES
B. S. Wilkins, MRCPath, and J.H. Williams, FlMLS 0 University Department of Pathology, Level E, South Block, Southampton General Hospital, Tremona Road, Southampton, SO9 4XY United Kingdom
J.A. Kohler, MRCP
Downloaded by [] at 15:44 22 March 2016
0 University Department of Child Health, Level E, South Block, Southampton General Hospital, Tremona Road, Southampton, SO9 4XY United Kingdom
D. B. Jones, PhD 0 University Department of Pathology, Level E, South Block, Southampton General Hospital, Tremona Road, Southampton, SO9 4XY United Kingdom
We performed an immunohistochemical analysis of frozen sections from testicular biopsies from 23 children with acute lymphoblastic leukemia. Eleven cases were infiltrated by leukemia. Tumor cells were immunostained by a panel o j antibodies that identified CDI 0, CD43, CDI 9, CD3, CD7, and MHC class I and II. The immunoreactivily of normal testicular components was also studied. Normal testis showed no CDlO reactivi& Wide variation in the number of stromal macrophages identified by CDllc was found. Fansjerrin receptor (CD71) was expressed by some stromal macrophages, by seminifnus tubules, and by Leydig cells. B lymphocytes were absent from the testicular stroma but small numbers of T lymphocytes were consistently present. MHC class I and II were expressed by most stromal cells but not by seminiferous tubules. KEY WORDS: acute lpphobkutic leukemia, immunostaining, testis.
INTRODUCTION The testis is a frequent site of involvement by acute lymphoblastic leukemia (ALL) in childhood, with a reported incidence of relapse at this site, during or after initial chemotherapy, of 8-16%.''3 These reports have also suggested that as a result of longer survival of children with acute leukemia, the frequency of testicular relapse is increasing.
Pediatric Hmtology and Oncology, 9:297-307, 1992 Copyright 0 1992 by Hnnisphne Publishing Corporation
297
Downloaded by [] at 15:44 22 March 2016
298
6. S. WlLKlNS ET AL
Testicular involvement may be an isolated finding,"3 but often heralds systemic relapse and thus warrants aggressive therapy even if evidence of systemic disease is absent . I Consequently, bilateral testicular biopsies have been performed as part of the assessment of male children after completing 2 years of maintenance chemotherapy. The aim has been to detect testicular involvement by leukemia at a subclinical stage, in the belief that early treatment would lead to improved survival. The need for routine testicular biopsies has been questioned in recent years, but the issue remains contr~versial.~-' In general, simple light microscopic analysis of biopsy morphology has been used to assess the presence or absence of leukemic infiltration within testes. The significance of diffuse or focal minimal infiltration has not been specifically addressed. It has been suggested that immunohistochemistry may be a useful diagnostic adjunct to routine histochemical staining when morphological assessment of the testicular biopsies is equivocal. Few studies have reported immunophenotypic data for leukemic testicular infiltrates1'"' and there has not been a systematic analysis of the immunoreactivities of normal testicular components, particularly of stromal cells that could cause diagnostic confusion with leukemic cells. We have reviewed immunohistochemical staining of testicular biopsies from children with ALL in order to characterize the antigen expression of leukemic infiltrates and normal testicular cells. We have also considered the value of immunostaining for detection of minimal involvement.
PATIENTS, MATERIALS, AND METHODS From 1985 to 1990, 29 boys with acute leukemia underwent testicular biopsy at Southampton General Hospital, either for symptomatic testicular enlargement (7 cases) or at the end of 2 years of maintenance chemotherapy (22 cases). Their ages at diagnosis ranged from 7 months to 14 years (mean age 6 years and 4 months). All of the children were treated according to Medical Research Council (MRC) protocols in use at the time of their illness. Fresh tissue was available from 24 of the cases, including 11 of 12 with infiltration by ALL. Immunostaining was performed using a well-established indirect immunoperoxidase technique" on frozen sections of each sample at the time of biopsy. All cases were studied using a diagnostic panel of monoclonal antibodies directed against a range of lineage-specific/lineage-related antigens. The antibody panel varied over the 5-year period; the antigens studied are described in the Results section, classified according to the proceedings 13 of the Fourth International Leukocyte Typing Conference. The immunophenotypes of the leukemic infiltrates were reviewed and the reactivities of normal testicular components with the antisera determined. The
TESTICULAR IMMUNOSTAININC IN ALL
299
subsequent clinical progress of patients in whom there was testicular involvement was also reviewed.
RESULTS
Downloaded by [] at 15:44 22 March 2016
Prevalence of Testicular Involvement by Leukemia Testicular involvement by leukemia was found in all 7 of the symptomatic and in 6 of the routine cases. One of the patients with testicular enlargement had a diagnosis of acute monoblastic leukemia and is not included hereafter. The other 12 boys had diagnoses of ALL (8 common ALL [CALL],2 T-ALL [T cell ALL], 2 ALL not otherwise specified); fresh tissue for immunostaining was available from 11 of these 12 boys. The twelfth child had cALL with subclinical testicular involvement at the time of routine biopsy. Table 1 summarizes the clinical course of each of the 11 ALL patients for whom immunostaining was performed and shows the results of clinical, histological, and immunohistochemical evaluation of testicular involvement. In two patients occult testicular involvement was found by immunostainTABLE
1. Diagnosis and Clinical Outcome of Patients in This Study
Testicular Involvement Diagnosis
C
H
I
Clinical Coursea
1
cALL
+
2
cALL cALL
4
T-ALL T-ALL ALL cALL
+ + + + + + + + + +
?T relapse at 2 y (no Rx). BM relapse at 3 y and T at 6 y. Further T relapse at 9 y followed by BM-Death. BM, T, and CNS relapse at 2 y. Autografted. Further BM
3
+ + + + + + + + -
Case
5 6 7
8 9 10
cALL CALL cALL
11
ALL
Abbreviations: C
+ + + + +
-
-
- -
+
relapse- Death. BM relapse at 4 y, T at 5 y, and further BM relapse at 6 y Death. T relapse at 1 y, BM relapse at 2 y-Death T relapse at 2 y. BM and CNS relapse at 3 y-Death. T relapses at 3 y and 4 y. Currently in remission at 9 y. BM and T relapse at 2 y. Further BM relapse at 2 y 6 mo- Death. BM and T relapse at 2 y. Allografted. Well at 5 y. T and CNS relapse at 2 y. BM relapse at 3 y-Death. Minimal T involvement at 2 y. Well at 2 y 8 mo. Autografted at 8 y but died. CNS relapse at 3 y. Minimal T involvement at 6 y. BM relapse at 6 y 6 mo. Autografted at 8 y but died.
-
-
-
-
detected clinically (testicular enlargement); H detected histologically in routine detected by immunohistochemical staining; T testicular; BM bone
H+E stained sections; I
marrow. 'Dates given as from the time of diagnosis.
300
B. S. WlLKlNS ET AL
ing alone. One of these patients (case 11) developed bone marrow relapse within 6 months of testicular biopsy; the other (case 10) remains well 8 months from the end of his chemotherapy, One further patient (case 3) had immunohistochemical evidence of leukemia in both testes despite clinical symptoms affecting one testis only.
Downloaded by [] at 15:44 22 March 2016
Distribution of ALL Infiltrates Within the Testis Patterns of leukemic infiltration within testicular stroma are shown in Figures 1-3. Infiltration was diffuse in most case (Fig. l ) , with sparing of tubules, but two cases showed marked leukemic infiltration of tubular basement membranes (cases 2 and 4; Fig. 2). Case 1 showed a tendency for localization of the infiltrate to perivascular areas (Fig. 3) in each of two episodes of unequivocal testicular relapse separated by an interval of 3 years. lmmunohistochemicalReactivities
of ALL Infiltrates Table 2 summarizes the tumor cell immunophenotypes of the positive ALL cases. All cases expressed CD10. Individual cases of CALL were also
FIGURE 1. Diffuse interstitial infiltration of testicular stroma by leukemic cells. Only occasional leukemic cells encroach on tubules. Immunostained for CD10.
Downloaded by [] at 15:44 22 March 2016
TESTICULAR IMMUNOSTAINING IN ALL
301
FIGURE 2. Diffuse interstitial infiltration by leukemic cells with extension into the layers of the basement membrane of seminiferous tubules. Immunostained for CDIO.
FIGURE 3. Perivascular condensation of leukemic cells within testicular stroma. Immunostained for CD10.
Downloaded by [] at 15:44 22 March 2016
302
B. S. WlLKlNS ET AL
variously positive for CD22, CD19, CD43, and MHC class I and 11. No expression of immunoglobulin light or heavy chains was seen. The two TALL cases reacted only weakly with CDlO but showed strong expression of CD3 and CD43. Immunostaining for terminal deoxynucleotidyltransferase (TdT) was not generally performed on our cases because the antibody produced only a weak reaction when positive and showed widespread nonspecific staining of all testicular components. Cell proliferation within the tumor infiltrates, as indicated by reactivity with Ki67, differed considerably between cases and did not appear to correlate with the clinical course. Case 1 suffered two episodes of symptomatic testicular relapse late after initial chemotherapy and biopsies were immunostained on both occasions. Between relapses further intensive chemotherapy was given. In the first biopsy 25% of cells were reactive with Ki67, while in the second biopsy this figure was reduced to 1%. In case 11 a microscopic focus of leukemic cells, not apparent in H&E stained sections, was detected by its reactivity for CDlO (Fig. 4) and in case 10 occasional CDlO reactive cells were scattered throughout the stroma. No other positive immunohistochemical marking was obtained in these cases. The occult infiltrate in the clinically normal testis in case 3 had the same immunophenotype as that of the leukemic cells within the abnormal one. In nine cases detailed immunophenotypic data for leukemic cells from the TABLE 2. Immunophenotypes of Leukemic Infiltrates Within Testicular Biopsies Antigen Diagnosis
CDlO
CD3
CD5
CD7
CD19
CD22
CD37
CD43
MHC-I
MHC-I1 ~~
CALL T-ALL ALL nos
717 212 212
015 212 112
014 212 nt
013 212
415
011
nt
111
416 nt nt
216 011
nt
215 212 nt
414 nt nt
214 nt 011
Results are expressed as number of cases positivelnumber of cases tested. No expression of immunoglobulin light or heavy chains was found. Cells in cycle, identified using the monoclonal antibody Ki67, varied from 1 % to 75%. Expression of CD2, 4, 8, 35, 38, and 71 was also studied in a proportion of cases. Both T-ALL cases expressed CD2 and case 5 was also positive for CD4, 8, and 71 (case 4 could not be tested for these antigens due to the small biopsy size). None of the other cases tested showed expression of these antigens. Antibodies used: 55, RFAL 1, 2, and 3 (CD10); OKT3, UCHTl (CD3); OKTl (CD5); HB2 (CD7); HD37 (CD19); To15 SCHLl, HD6 (CD22); WR17 (CD37); WRl4 (CD43); HB95 (MHC I, American Type Culture Collection); 185 (MHC 11-ICRF, Loncoln’s Inn Fields, London); Ki67 Gerdes. Forschungsinstitiit, Borstel, Germany); O K T l l (CD2); OKT4, 73F11 (CD4); OKT8 (CD8); El 1 (CD35); AT13.5 (CD38); HB21 (CD71). Abbreviations: C D cluster of differentiation. CD reactivities of individual antibodies are categorized according to the proceedings of the Fourth Leukocyte Typing Conference (see Ref. 13). MHC I and I1 major histocompatibility antigens, classes I and 11. nt not tested. nos not otherwise specified.
u.
-
-
-
-
Downloaded by [] at 15:44 22 March 2016
TESTICULAR IMMUNOSTAINING IN ALL
303
FIGURE 4. Microscopic focus of CD10-positive cells in the testicular stroma of case 11.
patients' blood or bone marrow at diagnosis were available for comparison with the results of our testicular immunostaining. Although there was only partial overlap between the panels of test antisera, no inconsistencies of antigen expression were found between leukemic cells at diagnosis and at testicular relapse. lmmunohistochemical Reactivities of Normal Testicular Components
Immunostaining of uninvolved testicular biopsies showed that Sertoli cells, gonocytes, stromal macrophages, and Leydig cells express CD7 1 (Fig. 5). Sertoli cells and endothelium react with antibody directed against CD38, an antigen expressed by late B lymphocytes and plasma cells (Fig. 6). One previous report has suggested that germ cells express CD38" but in our study they were negative. Most stromal cells express MHC class I and 11. Strong M H C class I expression by endothelial cells was also seen. The number of stromal macrophages, as identified by expression of CD1 lc, varied widely between biopsies. No B lymphocytes were identified in the normal stroma, but small numbers of T lymphocytes were consistently present. CD4- and CD8-positive T lymphocytes were present in equal numbers in most cases.
Downloaded by [] at 15:44 22 March 2016
304
6. S. WlLKlNS ET A 1
FIGURE 5. Stromal Leydig cells and cells within the seminiferous tubules (Sertoli cells and developing gonocytes) expressing CD71. Tubular staining is relatively weak, but comparison with the unstained seminiferous tubules of Figures 3 and 4 enables the cytoplasmic “blush” of CD71 to be appreciated.
DISCUSSION Testicular involvement by childhood ALL is a common accompaniment of the disease but in our series it was clinically apparent in only 6 of the 12 cases. In the other 6 cases leukemic infiltration was detected only by routine microscopy of hematoxylin-eosin (H&E)-stained sections and/or immunohistochemistry, including 2 patients in whom leukemic cells were detected only by immunostaining for CD10. One further patient who had clinical symptoms of leukemia involving only one testis showed immunohistochemical evidence of additional, sparse, leukemic infiltration of the other. The number of cases is too small for statistical analysis but there did not appear to be any correlation between the histological pattern or the extent of infiltration and the subtype of ALL present. The clinical significance of the presence of small number of CD10-positive cells within testicular stroma, as in cases 10 and 11, is unclear. CDlO is not uniquely expressed by ALL blasts, having also been demonstrated on some normal hematopoietic stem cells, renal and gut epithelium, neutrophils, melanoma and glioma cell lines, and cultured fibroblast^.""^ We saw no CDlO expression by normal testicular cells in 12 definitely uninvolved cases. Therefore it is unlikely that scattered CD10-positive cells in the testis simply repre-
Downloaded by [] at 15:44 22 March 2016
TESTICULAR IMMUNOSTAINING IN ALL
305
sent a normal stromal population and it seems reasonable to assume that they are leukemic cells. It remains a possibility that following intensive chemotherapy nonneoplastic, CD 10-positive, immature lymphoid cells could be present in tissues in response to inflammatory stimuli. Against this, we saw small numbers of phenotypically mature T cells in all of the biopsies that were not positive for leukemia, in the absence of any CD10-positive cells. Such difficulties in the interpretation of minimal or equivocal infiltration raise questions about the importance of detecting subclinical testicular relapse. Since the recommendation of routine testicular biopsies in the late 1970s, 3-9 many studies have attempted to assess the value of the procedure. Several have shown that clinically normal children with histological evidence of testicular relapse have a worse prognosis than those with normal histology, but few data are available to compare the outcome of clinically overt testicular relapse with relapse detected by histology alone. It has been suggested that it may be unnecessary to detect subclinical testicular relapse since current therapy permits successful reinduction of remission once the disease has subsequently become overt .6 Other workers, however, demonstrated a survival advantage associated with detection of testicular involvement at a subclinical rather than an overt stage3 or found that detection of occult testicular relapse identifies a group of patients who have a high risk of subsequent systemic relapse.' The
FIGURE 6. Immature seminiferous tubules showing unstained gonocytes (arrowed) among CD38positive Sertoli cells. The positive staining can be appreciated best by comparison with unstained tubules in Figures 3 and 4.
Downloaded by [] at 15:44 22 March 2016
306
B. S. WlLKlNS ET A1
value of performing testicular biopsies in asymptomatic boys at the end of maintenance chemotherapy remains an issue of controversy. The problem of interpreting the occasional finding of minimal or equivocal infiltration has been noted in several of the studies mentioned above, but the outcome for patients with such findings is not known. Clinical outcome for the CALL cases in our study was variable and the numbers were too small to indicate whether specific immunophenotypes had good or bad prognostic significance. The two T-ALL cases had rapid progression of their disease, consistent with their having very high white blood cell counts at diagnosis (data not shown). In both cases testicular relapse was clinically obvious before completion of maintenance therapy. Variation in Ki67 reactivity may be an intrinsic feature of the disease in each patient and/or an effect of therapy. That the latter is contributory is supported by case 1, in whom there was a substantial reduction in cycling (Ki67-positive) cells between the two biopsies from different episodes of testicular relapse. As no normal testicular components were found to express CD10, we consider that this antigen is the most reliable of currently detectable markers for ALL infiltrates within the testis. Other antigens that might identify leukemic cells are expressed by normal stroma (eg, various T-lymphocyte- and macrophage-associated antigens, MHC class I and II), tubules, and Leydig cells (CD71). Our finding of CD38 expression by Sertoli cells and not by gonocytes would appear to be at variance with results obtained by Verdi et al," who reported that germ cells were CD38-positive. The photomicrograph with which they illustrate CD38 immunostaining strongly suggests that Sertoli cells were positive, although they make no reference to Sertoli cell reactivity in their text. The seminiferous tubules are not shown at sufficiently high magnification to identify germ cells with uncertainty, and the discrepancy may lie in the terminology used to describe tubular cells rather than in a true difference in reactivity between our two studies. B lymphocytes were not found in uninvolved testicular stroma, in keeping with our previous findings in the adult testis." As some B-cell-associated antigens (eg, CDl9 and CD22) were commonly expressed by ALL cells, these are useful additional indicators of leukemic infiltration.
REFERENCES 1. Stoffel TJ, Nesbit ME, Seymour HL. Extramedullary involvement of the testes in childhood leukemia.
Cnncn. 1975;35:1203-1211. 2 . Hustu HO, Aur RJA. Extramedullary leukaemia. Clin Hasmafol. 1978;7:313-337. 3. Bowman WP, Aur RJA, Hustu HO, Rivera G. Isolated testicular relapse in acute lymphocytic leukemia of childhood: categories and influence on survival. J Clin Oncol. 1984;2:924-929.
Downloaded by [] at 15:44 22 March 2016
TESTICULAR IMMUNOSTAINING IN ALL
307
4. Askin FB, Land VJ, Sullivan MP, et al. Occult testicular leukemia: testicular biopsy at three years continuous complete remission of childhood leukemia: a Southwest Oncology Group study. Cuncn. 1981;47:470-475. 5. Nachman J, Palmer NF, Sather HN et al. Open-wedge testicular biopsy in childhood acute lymphoblastic leukemia after two years maintenance therapy: diagnostic accuracy and influence on outcomea report from Children’s Cancer Study Group. Blood. 1990:75:1051-1055. 6. Hudson MM, Frankel LS, Mullins J, Swanson DA. Diagnostic value of surgical testicular biopsy after therapy for acute lymphocytic leukemia. J Pediatr. 1985:107:50-53. 7. Wong KY, Ballard ET, Strayer FH, et al. Clinical and occult testicular leukemia in long-term survivors of acute lymphoblastic leukemia. J Pediutr. 1980;96:569-574. 8. Ortega JJ, Javier G, Toran N. Testicular infiltrates in children with acute lymphoblastic leukemia: a prospective study. Med Pediatr Oncol. 1984;12:386-393. 9. Ise T, Kishi K,Imashuku S, et al. Testicular histology and function following long-term chemotherapy of acute leukemia in children and outcome of the patients who received testicular biopsy. Am J Pediutr H ~ m ~ t On~ol. ol 1986;8:288-293. 10. Verdi CJ, Hutter J, Grogan T M . Immunophenotyping to detect and characterize acute lymphocytic leukemia in testicular biopsies. Pediutr Padol. 1989;9: 117-130. 11. Wilkins BS, Williamson JMS, O’Brien CJ. Morphological and immunohistochemical study of testicular lymphoma. Histoputhology. 1989;15:147-156. 12. Jones DB, Moore K, Wright DH. Heterogeneity of B-cell antigen expression in non-Hodgkin’s lymphoma. Adv Exp Med Biol. 1989;137:139-144. 13. Knapp W, et al. (eds). Leukocyte Typing, Vol. 4. Oxford: Oxford University Press;1989. 14. Hoffman-Fezer G, Knapp W, Thierfelder S. Anatomical distribution of CALL antigen expressing cells in normal lymphatic tissue and in lymphomas. Leukemia Res. 1982;6:761-767. 15. Dorken B, Moller P, Pezzutto A, et al. B cell antigens: CDIO. In: Knapp W, et al, eds. Leukocyte Typing, Vol. 4. Oxford: Oxford University Press;1989:33-34.
Received October 14, 1991 AcceptedJunuury 22, 1992
