- Email: [email protected]
DNA in Non Hodgkin-Lymphoma —A Cytophotometric Study
Beitr. Path. Vol. 160,213-230 (1977)
Original Papers
Department of Pathology (Ludwig-Aschoff-Haus), University of Freiburg (Germany)
DNA in Non Hodgkin-Lymphoma A Cytophotometric Study W. SANDRITTER and H. GRIMM
Summary Cytophotometric DNA determinations on 26 cases of non-Hodgkin's lymphoma yielded the following findings: r. Follicular centrocyticicentroblastic lymphomas (M. Brill-Symmers) and diffuse centro-
cytic lymphomas (lymphocytic lymphosarcoma) have a diploid DNA stem line. Diploid DNA values are observed in benign tumors, so that the assignment of these lymphomas to the group of "low grade malignancies" appears justified. 2.
Lymphoblastic sarcomas show an aneuploid DNA stem line, as do 96% of all malignant tumors.
3. Lymphoid cells, plasma cells, and immunoblasts seen in immunocytomas are aneuploid. Thus these lymphomas must belong to the group of "high-grade malignant lymphomas" as regards their DNA distribution. 4. Immunoblastic sarcomas have aneuploid DNA stem lines (r case tetraploid), in which both the lymphoid cells and the plasma cells from those immunoblastic sarcomas arising from immunocytomas show atypical DNA distribution patterns. 5. In two cases of angioimmunoblastic lymphadenopathy, the lymphoid cells, plasma cells, and immunoblasts arc aneuploid. They are thus regarded as "high grade malignancy" lymphomas. The results are discussed with respect to climcal course and prognosis. Measurements on a larger series of cases and correlation to clinical data are needed to support these results. Ultrafast DNA measurements made by flow-through cytophotometry can perhaps be helpful in the future for making the decision between a "low" or "high" grade malignant lymphoma.
Significant controversy surrounds the classification of the non-Hodgkin's lymphomas (malignant lymphomas, ML). The principal classification employed in English-speaking countries is that of Rappaport et al. (1956), added to which the World Health Organization (WHO) has submitted its own suggestions (Mathe et aI., 1976), and the classification by Lennert 15 Beitr. Path. Vol. 160
2
14 . W. Sandritter and H. Grimm
(the so-called "Kiel Classification", Gerard-Marchant et aI., 1974) has been under discussion since I 974. A classification of non-Hodgkin's lymphomas should not simply satisfy academic interests, but should involve prognostic perspectives as well. The Kiel classification satisfies this requirement - it identifies low grade malignancy MLs and high grade malignancy MLs - and numerous studies have appeared showing the clinical course of cases classified according to Rappaport. The goal of the following study is to contribute to the classification and to the problem of prognosis in MLs on the basis of cytophotometric DNA measurements. With regard to classification, we would hope to decide in some cases whether we are dealing with a uniform or a mixed cell population. With regard to prognosis, we are supported by our earlier investigations (B6hm and Sandritter, 1975) which have shown that benign tumors always exhibit a diploid DNA distribution pattern whereas 96% of malignant tumors show aneuploid frequency peaks. The DNA di~tribution patterns in 26 non-Hodgkin's lymphomas were investigated with respect to these two questions.
Materials and Methods Imprints were made from formalin-fixed lymph node biopsies. Lymphocytes from cases of non-specific lymphadenitis served as reference cases for the diploid, 2C DNA value. Cases were fixed for varying lengths of time in formalin. The imprints were fixed for one hour in a solution containing methanol 85%, formalin 10%, and glacial acetic acid 5%. Hydrolysis for the Feulgen reaction was performed in 4 N HCI for 60 min at 28° C. After the usual Feulgen procedure (Grauman, 1953), the slides were mounted in Cargille refraction fluid, nD 1.54. Cytophotometric measurements were made on an integrating microdensitometer (Barr and Stroud, Glasgow). The measuring point diameter was 0.5 !Am, the numerical aperture of the condenser was 0.3, and the numerical aperture of the objective was 1.3. The resulting measurements of DNA content are given in arbitrary units (AU). Between 100 and 200 cells were measured for each case. Error: Measurements made in ten diploid liver cells resulted in a standard deviation of ± 3%. The overall error, involving fixation, staining, and measurement errors, is estimated at ± 15-20%. Deviations from the diploid lymphocyte value (2C) were only counted as "true" deviations when they lay outside this 20% range.
Results Twenty-six cases of non-Hodgkin's lymphoma were measured. Diagnostic assignments for each case were made according to the old nomen-
DNA in Non Hodgkin-Lymphoma.
215
clature, the nomenclature of Lennert (Kiel Classification, 1974), and according to the classification of Rappaport et al. (1956)1). M. Brill-Symmers (ML follicular, centrocytic centroblastic; ML nodular, poorly differentiated) 1.
In all four cases studied, we were dealing with typical, follicular lymphomas with a predominant growth of germinocytes (centrocytes). Germinoblasts were present only sparsely. Both cell types were readily distinguished in the imprint preparations, and were measured separately. Figure I shows the DNA distribution pattern of a typical case (15624/75), which is representative for all cases. We see here that the DNA value for germinocytes (centrocytes) is very close to that for lymphocytes, and shows only a limited scatter toward the right up to 26 AU (diploid 18 AU, germinocyte peak at 22 AU). The germinoblasts (centroblasts) have values in the diploid region to a lesser extent, and scatter very strongly over the tetraploid value (S-phase cells) without a distinct peak formation. Table I shows that the same distribution holds for the other cases: the germinocytes are diploid with limited scatter, while the germinoblasts exhibit both diploid and interdiploid to tetraploid values. Two cases showed a hyperdiploid (triploid) peak formation.
Table 1. M. Brill-Symmers (ML follicular, centrocytic, centroblastic; ML nodular, poorly differentiated) Nr.
Age
Sex
DNA distribution pattern
15 62 4/7 6
63
~
Germinocytes diploid. Germinoblasts diploid, with scatter exceeding tetraploid (Fig. I).
16164/75
65
((
Germinocytes diploid. Germinoblasts diploid to tetraploid.
18 457/75
72
0
Germinocytes diploid. Germinoblasts triploid.
29 12 4/75
66
<3
Germinocytes diploid to triploid. Germinoblasts triploid with scatter to tetraploid. - --~--------~
1) We are very grateful to Prof. K. Lennert and Prof. H. Rappaport for reviewing our material.
216 . W. Sand ritter and H. Grimm 42
n 32 _ 28 24 _
15624/75
....
M. Brill- Symmers
20 _ 16 12 _ 8 _
Germinocyles
.
!,.. u 0
~!4
,..
E
-J
P'"
I
116t20
30
t
40
50
AU
60
20
Germinoblasls 16 12 8 4
30
t
40
50
60
Fig. I. DNA distribution pattern for germinocytes and germinoblasts in a case of M. Brill-Symmers (I5624i 75). Lymphocytes 2C = diploid = 18 AU, first arrow. Tetraploid ,= 4C = 36 AU, second arrow. n = number of cells. AU = relative DNA content in arbitrary units.
2. Lymphocytic lymphosarcoma (ML centrocytic, ML diffuse, poorly differentiated) The four cases studied showed the typical histological picture of a lymphocytic lymphosarcoma, in which the cells exhibit the appearance of centrocytic cells with very angular nuclei and an easily folded nuclear membrane without visible cytoplasm (see Lennert et al., 1975a, Fig. 3). Figure 2 (26507/7 5) gives a typical DNA distribution pattern with a peak at 30 to 32 AU (lymphocytes 28 AU). The peak of the DNA values thus lies within the variation of the diploid value. The limited scatter to higher values can be attributed to S-phase cells. All other cases showed practically the same DNA distribution pattern, without a clear peak formation in the
DNA in Non Hodgkin-Lymphoma.
26507/75
21 7
Lymphocytic Lymphosarcoma
n
24 20 16
12
8 4
40
50
t
70
60
AU
Fig. 2. DNA histogram for a lymphocytic lymphosarcoma (26507/75). Diploid AU, first arrow. Tetraploid = 56 AU, second arrow.
28
tetraploid region. Case 26507/75 had a limited amount of sclerosis, so that the lymphocytic, diffuse form of lymphosarcoma as described by Bennett and Millett (1969) is present. The patient expired five months after the initial diagnosis was made on the basis of a cervical lymph node, with autopsy findings of a generalized lymphosarcoma: lymph nodes, stomach, kidney, retroperitoneum, heart, bone.
Table ated)
2.
Lymphocytic lymphosarcoma (ML centrocytic; ML diffuse, poorly differenti-
Nr.
Age
26 50 7/75 10 301 /75
54 68
120 51/75
57
28 94 0 /75
44
Sex
DNA distribution pattern
o
Diploid with scat
o o o
'~r
to triploid values (Fig. 2).
Diploid with scatter to triploid values. Diploid. Diploid with scatter to tetraploid values.
3. Lymphoblastic lymphosarcoma (ML lymphoblastic or undifferentiated) This group of tumors is readily distinguished in terms of its DNA distribution pattern from the lymphocytic lymphosarcomas. All four cases show peak formation at hyperdiploid or triploid values.
2 18 •
Table 3.
W. Sandritter and H. Grimm Lymphoblastic lymphosarcoma (ML lymphoblastic)
Nr.
Diagnosis
Age
Sex
DNA distribution pattern 1)
353 8/7 6
Burkitt
I2
0
Hyperdiploid. Peak 38 AU.
24 I 53/75
Lymphoblastic
I5
~
Hyperdiploid (Peak at 36 AU), second peak around 70 AU (Fig. 3).
1735?/75
Lymphoblastic
60
0
Hyperdiploid (Peak at 38 AU), second peak around 70 AU.
897/75
Lymphoblastic
68
~
Triploid (43 A), second peak around 84 AU (Fig. 4).
1) Diploid value 28 AU, tetraploid 56 AU, octoploid II2 AU.
Case 3538/76 involves a histologically typical Burkitt's lymphoma with a frequency peak of DNA values at 38 AU hyperdiploid and a wide scatter of values up to 40-60 without a second peak (Tab. 3). The remaining three cases (24153/75,897175,17357175) are unclassifiable lymphoblastic sarcomas (non-convoluted; after Rappaport: diffuse, lymphoblastic, undifferentiated). The DNA distribution for case 24153/75 shows a hyperdiploid peak (36 AU) with wide scatter up to an implicit doubling-value of
n 40
24153/75
Lymphoblastic
Lymphosarcoma
36 32 28 24 20 16 12 8 4
28
t
30
40
50
t
60
70
80
90
AU
Fig. 3. Lymphoblastic lymphosarcoma (24153/75) with a hyperdiploid (36 AU) DNA stem line (First arrow diploid value; second arrow tetraploid).
DNA in Non Hodgkin-Lymphoma. 219
n
897/75 Lymphobl..tlc Lymphosarcoma
40 36 32
28 24 20 16
12
8
100
120
AU
Fig. 4. Lymphoblastic lymphosarcoma with a triploid DNA stem line (42 AU) and a second peak at circa 80 AU (hypooctoploid). First arrow = tetraploid value. Second arrow = octoploid value.
70 AU (hypertetraploid) (see Fig. 3). Case 17357/75 shows the same pattern (peak at 38 AU with a weak doubling peak), and in case 897/75 (Fig. 4) we find a triploid peak (42 AU with doubling at 80 AU). Cells in GI phase (hyperdiploid or triploid) were distinguishable from cells in G2 phase (in the doubling peak) in all three cases by means of a larger nucleus and more open chromatin pattern.
4. Immunocytoma (ML lympho-plasmocytic-plasmacytoid) Both we and Lennert made the diagnosis of lympho-plasmacytoid immunocytoma for these four cases (829/75 and 2929I!75 polymorphic type). Rappaport made varying diagnoses (see Tab. 4), where it should be recognized that methyl-green pyronine slides were not available to Rappaport. All cases showed the typical, histological picture: diffuse infiltration with lymphoid cells which were not further classifiable, small lymphoid plasma cells (typical as well as large, atypical plasma cells), and large immunoblasts with basophilic cytoplasm and 1-2 large nucleoli (see Lennert et al., 1975b, Fig. 9). The various cell types were distinguishable only to a limited extent on the basis of nuclear morphology in the Feulgen-stained smear preparations. Our cytophotometric measurements were thus restricted to distinguishing between lymphoid cells and small plasma cells on the one hand and cells with large nuclei (large plasma cells and immunoblasts) on the other.
130
220 . W. Sandritter and H. Grimm Table 4.
lmmunocytoma (ML Iympho-plasocytic-plasmocytoid; ML !'p.)
~-------~-,------.
No.
Diagnosis
Age
851 0 /7 6
ML!.p
54
769/75
ML !.p2)
82 9/75
ML !'p3)
29 2 9
r175
Sex
DNA-distribution pattern l ) Hyperdiploid 38-40 AU and hypertetraploid (around 60 AU) (Fig. 5). Hyperdiploid 38 AU and hypertetraploid 60-70 AU (Fig. 6).
54
ML !.p4)
Hypotetraploid 44 AU. Hypooctoploid 86 AU. Hyperdiploid 34 AU scattered values at 76 AU.
- _ . _.._----- -- ----------
1) 2} 3) 4)
Diploid Value 28 AU, tetraploid 56 AU. After Rappaport: ML diffuse, well differentiated, lymphocytic. After Rappaport: ML diffuse, mixed. After Rappaport: ML diffuse, well differentiated, lymphocytic. 4) After Lennert: ML Iympho-plasmocytoid, polymorphic. ad 2) 769/75 Autopsy findings: tumor infiltrates in all lymph nodes, skin, stomach, gut, liver, kidneys, spleen, bone marrow. Course: 6 years. ad 3) 829/75 Autopsy findings: para aortic lamph nodes, peritoneum. Course: 6 months. Polymorphic type.
Table 4 shows that the DNA distribution pattern is rather similar for each of the four cases (see Figs 5 and 6). The DNA values for the small lymphoid cells and lymphoid plasma cells show their first frequency peak in the hyperdiploid region (hypotetraploid for case 829/75), while the large nucleus cells (atypical or typical plasma cells, immunoblasts) exhibit a second frequency peak at 60-70 AU (hypertetraploid).
5. I mmunoblastic sarcomas This group of tumors has had a variety of designations: reticulum cell sarcoma (Mathe et aI., I976), ML diffuse histiocytic (Rappaport et aI., I956), and immunoblastic sarcoma (Lennert, Kiel Classification). Our 8 cases consist in part of pure immunoblastic sarcomas, as read by Rappaport, which, however, were diagnosed as malignant lymphomas of the centroblastic type by Lennert (see Tab. 5). Four cases involved immunoblastic sarcomas arising from an immunocytoma, so that they are rich in lymphoid cells, plasma cells, and small and large immunoblasts. Figure 7 gives an example of the DNA distribution pattern of the immunoblastic sarcomas arising from an immunocytoma (23395/75). The lymphoid cell elements and the small immunoblasts show widely scattered
DNA in Non Hodgkin-Lymphoma·
221
n 40 36
8510/76
Immunocytoma
32 28 24 20 16 12 8 4
t
20
30
t
50
40
60
70
AU
Fig. 5. Immunocytoma (8510/75) with a hyperdiploid DNA stem line (38-40 AU) (lymphoid cells) and a second peak in the hypertetraploid region (60-70 AU). First arrow = diploid. Second arrow = tetraploid.
n
769/75
Immunocytoma
40 36 32 28
20 16
12
30
40
50
t
60
70
80
90
100
120
AU
Fig. 6. Immunocytoma (769/75) with a hyperdiploid DNA stem line (38 AU) (lymphoid cells) and hypertetraploid DNA values (60-70 AU). Wide scatter of values up to 130 AU (112 AU = octoploid). First arrow = tetraploid value. Second arrow = octoploid value.
130
222 .
W. Sandritter and H . Grimm
n 40
23395/75
Immunoblast ic Sarcoma
36 32
28 24 20 16 12 8 4
t
30
40
50
100
AU
Fig. 7. Immunoblastic sarcoma arising from an immunocytoma (23395 /75). Lymphoid cells and small immunoblasts show values in the triploid to hypotetraploid region 40-50 AU). Second peak at 80 AU. First arrow = diploid value. Second arrow = tetraploid value.
n
28091/7'
Immunoblestic Sarcoma
32 28 24 20 1&
12 8
30
40
50
t
eo
70
80
90
100
Fig. 8. Immunoblastic sarcoma (2609I /75) with a tetraploid DNA stem line (56 AU) and wide scatter up into the octoploid region (I I2 AU). First arrow = tetraploid value. Second arrow = octoploid value.
DNA in Non Hodgkin-Lymphoma. 223
values (40-50 AU = triploid-hypotetraploid), and a doubling peak is seen at 80 AU (Fig. 7). All other cases show practically the same DNA distribution pattern, with few variations (see Tab. 5). Figures 8 and 9 shows immunoblastic sarcomas without lymphoid cells, having only immunoblasts (small immunoblasts in GI phase and large immunoblasts in G2 phase) with basophilic cytoplasm and centrally positioned nucleoli in a clear nucleus. Case 2609ri75 (Fig. 8) is unambiguously tetraploid with a weak doubling peak at IOO-I2O AU. Case 18082/75 (Fig. 9) shows triploid values with wide scatter and no second peak formation. The two remaining cases behave in a similar fashion.
Table 5. No.
Immunoblastic Sarcomas (IS) (Reticulosarcoma, ML diffuse, histiocytic) Diagnosis
Age
Sex
DNA Distribution pattern 1)
- - - - - - ---- - - - -
IS arising from an immunocytoma (spleen, liver)
67
Lymphoid cells and small immunoblasts with triploid to hypotetraploid values (40-50 AU). Doubling peak at 80 AU (Fig. 7).
IS from immunocytoma
75
Lymphoid cells with triploid values. Small immunoblasts with hypotetraploid values. No clear, second peak. Values from 60 to 132 AU.
IS from immunocytoma
7I
Hyperdiploid (36 AU). No clear second peak. Scatter to 76 AU.
IS from immunocytoma (stomach)
73
Lymphoid cells and small immunoblasts with hyperdiploid values (36 AU). Immunoblasts peak at 72 AU, wide scatter.
IS2)
71
Tetraploid (56 AU). Values at 100-130 AU (Fig. 8).
18082/75
IS3)
35
206 38/75
IS4)
53
o o
Triploid (42 AU). Marked scatter (Fig.9). Tripoloid (42 AU). Values between 70 and 82 AU. Hyperdiploid (36 AU). Wide scatter, no second peak.
---~----~~---
1) Diploid Values 28 AU, tetraploid Values 56 AU. 2) Our diagnosis: immunoblastic sarcoma. Rappaport: ML diffuse histiocytic. Lennert: ML centroblastic. 3) Same diagnosis as 2. -1) Same diagnosis as 2.
224 . W. Sand ritter and H. Grimm n 32 18082/75
Immunoblastlc Sarcoma
28 24 20 18 12
8 4
t
~o
70
80
80
90
100
AU
Fig. 9. Immunoblastic sarcoma (18082/ 75) with a triploid DNA stem line and wide scatter of DNA values. First arrow = diploid value. Second arrow = tetraploid value.
n 18977/75 36
Angiolmmunoblastlc Lymphadenopathy
32 28 24
20 18 12
8 4
1
30
40
50
i
60
70
80
90
AU
Fig. 10. Angioimmunoblastic lymphadenopathy (18977/ 75). Lymphoid cells and plasma cells are near-triploid, immunoblasts are hypertetraploid. First arrow = diploid value Second arrow = tetraploid value.
6. Angioimmunoblastic lymphadenopathy (Lymphogranulomatosis X) This lymphoma, described by Frizzera et al. (I974) and Lukes et al. (I975), had been designated as lymphogranulomatosis X by Lennert in
DNA in Non Hodgkin-Lymphoma .
225
n 40 36 32 28
19095/7S
24
Angioimmunoblastic
lymphadenopathy 20
lymphoid cells
16 12
8 4
30
40
so
30
40
50
t
AU
60
24 20 16 12 8
t
60
70
80
90
AU
Fig. I I. Angioimmunoblastic lymphadenopathy (19095 /75). Lymphoid cells are hyperd iploid. Immunoblasts are triploid (42 AU) . 28 AU = diploid value. Arrow = tetraploid value.
1973. Our two cases show a proliferation of venules and a mixed cellular infiltrate of lymphoid cells, plasma cells, immunoblasts, histiocytes, and giant cells, which can resemble Sternberg giant cells. The Feulgen preparations only allow us to distinguish lymphoid cells and plasma cells on the one hand from small and large immunoblasts on the other, where the latter are characterized by their looser chromatin pattern and a central nucleolus. The lymphoid cells and small immunoblasts show a near-triploid peak in case 18977/75, while the larger immunoblasts scatter stongly to higher values up to 70-80 AU (Fig. 10). The limited scatter in the small lymphoid cells is particularly noteworthy. Case 19095/75 shows a similar DNA distribution pattern (Fig. r r).
226 . W. Sand ritter and H. Grimm Table 6.
Angioimmunoblastic lymphadenopathy (AL) (Lymphogranulomatosis X) ~-~~~---~---
No.
Diagnosis
~----~--
Age
~-~----------
Sex
--~-~~~~-
DNA-Distribution Pattern Lymphoid cells, plasma cells with near triploid values (40 AU). Immunoblasts hypertetraploid (70-80 AU). Wide scatter (Fig. r 0).
79
AL2)
Lymphoid cells hyperdiploid (38 AU). Immunoblasts 40-42 AU. Wide scatter. (Fig. r r). -----~
-------~
- - -
1) After Lennert: Al or T-Zone Lymphoma. 2) Rich in Plasma cells.
Discussion Almost no studies have been performed on the DNA-distribution pattern of the non-Hodgkin's lymphomas. The above cytophotometric measurements made on 26 cases sought to answer the extent to which it is possible to judge the malignancy of these lymphomas on the basis of the DNA distribution pattern (euploidy, aneuploidy). We further sought to answer whether a given cell population was uniform or mixed. It is necessary to realize that for DNA values lying between diploid (2C) and tetraploid (4C), or between tetraploid and octoploid (8c), we cannot decide whether we are dealing with cells in normal S-phase as opposed to aneuploid cells. Furthermore, DNA content makes no predictions about chromosomal abnormalities. Since we had available only formalin fixed material instead of fresh imprints, we were dealing with an error of about the diploid value of 15-20%. Finally, it is known from numerous prior investigations that lymphocytes often give 10% lower DNA values thaa, say, diploid liver cells (problem of chromatin condensation). Follicular lymphomas, centroblasticlcentrocytic (M. Brill-Symmers) showed two cell populations. Centrocytes were diploid (within the 15 to 20% range of scatter), while the centroblasts (germinoblasts) scattered up to tetraploid values. Wilms (1970) found the same distribution pattern in four cases of Brill-Symmers which were studied in Lennert's department. The interdiploid-tetraploid DNA values seen in centroblasts are presumably attributable to S-phase cells. Mitrou et al. (1969) found that normal germinoblasts from normal germinal centers showed 3H thymidine uptake in up to 50% of the nuclei, and the corresponding DNA contents for these cells lay in the interdiploid-tetraploid range. The corresponding germino-
DNA in Non Hodgkin-Lymphoma.
227
cytes showed only a very limited 3H thymidine uptake (8%). In this regard, we consider the rather indistinct peak formation of DNA values in the triploid range seen in two cases not as an aneuploid DAN stem line, but more likely a random accumulation of S-phase values (partial synchroni. .. ~) zatlon We conclude from the four cases we studied and from the measurements made by Wilms that these cases are characterized by an euploid distribution pattern, compatible with the relatively benign character of these lymphomas (low malignancy, according to the Kiel classification). Previous studies in our laboratory (Bohm and Sandritter, 1975) have shown that benign tumors are always euploid. Jones et al. (1973) also found a significantly higher survival rate in nodular lymphomas as compared to the diffuse lymphomas. Further studies on a larger number of cases will be necessary in order to reach a definitive conclusion on this matter 1). We found an entirely similar DNA distribution pattern for lymphocytic lymphosarcomas (ML diffuse centrocytic). All cases were diploid with only a limited scatter of DNA values (d. Hale et al., 1959). These entities are also rated as low-grade MLs by the Kiel Classification. Jones et al. (1973) found a lower survival rate for the diffuse malignant lymphomas. Musshoff et al. (1976) were unable to show a difference in the acturial survival rate for nodular and diffuse centrocytic malignant lymphomas. Lymphoblastic lymphosarcomas belong to the high grade MLs. Nathwani et al. (1976) found a mean survival rate of 8 months. All four cases in this study (one Burkitt's and three non-convoluted) showed an aneuploid DNA distribution. (DNA stem line in the hyperdiploid or triploid region, often with a hypertetraploid doubling peak). We found such aneuploid DNA stemlines only in malignant tumors in our ealier studies (Bohm and Sand ritter, I975). In other words, the DNA measurements corroborate the high grade malignancy of these MLs. The immunocytomas exhibit a mixed cell population, for which it is difficult to distinguish the individual components in the Feulgen-stained preparations. Our DNA measurements show that the immunocytomas consist of aneuploid cells as do other malignant tumors (Bohm and Sandritter, 1975), containing plasma cell, lymphoid cell, and immunoblastic components. The extent to which the cells in the first and second frequency peaks 1) The results of our observations are only partially in agreement with the DNA determinations of R. Silvestrini et al. (J. Nat. Cancer Inst. 58, 499, 1977) on nonHodgkin's lymphoma. In M. L. immunoblastic, lymphoblastic and immunocytoma, as well as in M. L. centrocyticlcentroblastic, these workers found diploid DNA values. At the moment we are unable to explain this discrepancy.
228 .
W. Sandritter and H. Grimm
(hyperdiploid, hypertetraploid) arise at least in part from a common cell population cannot be decided from the measurements. We suspect that small immunoblasts are hidden among the small lymphoid cells, and that these immunoblasts form the G2 cells in the second peak, but this second peak could equally well consist of plasma cells. Immunocytomas are regarded as low grade malignancy MLs after the Kiel Classification. We have two cases with good follow-up information: one case (769/75) survived 6 years, the other (829/ 75) survived 6 Months. Musshoff et al. (1976) found a better survival rate for diffuse lymphoplasmocytoid immunocytomas, but this is not statistically different from that of the high grade malignant lymphomas. The DNA measurements (aneuploid DNA distribution pattern with DNA stem lines) assign the immunocytomas to the group of high grade malignant lymphomas. Further studies will be necessary before reaching a final judgment on this matter. The transformation to an immunoblastic sarcoma is of particular interest, especially arising from a polymorphic immunocytoma. The immunoblastic sarcomas are morphologically a heterogeneous group. Immunoblastic sarcomas which arise from immunocytomas show typically hyperdiploid or triploid lymphoid cells (lympoid cell elements, plasma cells) and small and large immunoblasts which may have triploid or hypotetrapolid DNA-values with a doubling peak (hypooctoploid). We are obviously dealing with two aneuploid cell populations. Pure immunoblastic sarcomas may show hyperdiploid, triploid, or tetraploid DNA stem lines. Lennert diagnosed a centroblastic ML in three of four cases. The cytophotometric measurements make no distinction between immunoblastic sarcoma and centroblastic ML. The immunoblastic sarcomas were earlier regarded as reticulum cell sarcomas. We cannot be certain that true reticulum cell sarcomas do not arise in lymph nodes. Reticulum cell sarcomas of the skin seem to be exceptional in their clinical behavior as well (protracted course). Our own DNA measurements (Knoth et al., 1965) made on a reticulum cell sarcomatosis in the skin showed a tetraploid DNA stem line, as in the above case 26091 / 75· Two cases of angioimmunoblastic lymphoadenopathy (AL) were found to be cytophotometrically heterogeneous as well. Lymphoid cells and plasma cells showed a hyperdiploid or triploid DNA stem line, while the immunoblasts showed a wide scatter of values. Thus AL showed a DNA distribution pattern characteristic of a malignant tumor. Clinically, AL appears to be a highly malignant lymphoma, with a course of 1-18
DNA in Non Hodgkin-Lymphoma . 229
months (Lennert et al., 1974) or 1-45 months (Radaszkiewicz and Lennert, 1975). Further studies on a larger series of MLs must be done in order to show the extent to which cytophotometric DNA measurements, correlated with observations of clinical course, can offer a more accurate prognostic judgment. Lymph node material can also readily be worked up for DNA measurements by means of flow-through cytophotometers (pulse cytophotometry), so that a rapid evaluation with immediate diagnostic feedback becomes possible.
Zusammenfassung Zytophotometrische DNS-Bestimmungen an 26 Fallen von Non-Hodgkin-Lymphomen ergaben: 1.
Follikulare zentrozytischlzentroblastische Lymphome (M. Brill-Symmers) haben ebenso wie die diffusen zentrozytischen Lymphome (lymphozytisches Lymphosarkom) eine diploide DNS-Stammlinie. Diploide DNS-Werte werden in gutartigen Tumoren beobachtet, so dag die Einstufung dieser Lymphome in der Gruppe "low grade malignancy" gerechtfertigt erscheint.
2.
Lymphoblastische Sarkome zeigen eine aneuploide DNS-Stammlinie, wie 96 % malignen Tumorcn.
aller
3. Bei den Immunozytomen sind lymphoide Zellen und Plasmazellen ebenso wie die Immunoblasten aneuploid. Die Lymphome miifhen, nach der DNS-Verteilun~ zu urteilen, zu den "High-grade-malignancy" -Lymphomen zu rechnen sein. 4. Immunoblastische Sarkome zeigen aneuploide DNS-Stammlinien (I Fall tetraploid), wobei auch die lymphoiden Zellen und Plasmazellen bei den immunoblastischen Sarkomen, die aus Immunozytomen entstanden sind, atypische DNS-Verreilungsmuster aufweisen. 5. In zwei Fallen von angioimmunoblastischer Lymphadenopathie sind lymphoide Zellen, Plasmazellen und Immunoblasten aneuploid. Sie miissen demnach als "High-grademalignancy"-Lymphome angesehen werden. Die Ergebnisse werden im Zusammenhang mit dem klinischen Verlauf und der Prognose diskutiert. Messungen an einer grogeren Zahl von Fallen und Korrelation zu den klinischen Daten solhen diese Ergebnisse untermauern. Uhraschnelle DNS-Messungen mit der Durchflugphotometrie konnen vielleicht in Zukunft hilfreich sein fur die Entscheidung "low" oder "high malignancy" lymphom.
Literatur Bennett, M. H., and Millett, Y. L.: Nodular sclerotic lymphosarcoma. A possible new clinico-pathological entity. Clin. Radio!. 20, 339-343 (1969) Bohm, N., and Sandritter, W.: DNA in human tumors: A cytophotometric study. Curro Top. Path. 60, 152-214 (1975) (Springer-Verlag, Heidelberg) Frizzera, G., Moran, E. M., and Rappaport, H.: Angio-immunoblastic lymphadenopathy with dysproteinaemia. Lancet 1070-1073 (1974) 16 Beitr. Path. Vol. 160
230 . W. Sandritter and H. Grimm Gerard-Marchant, R., Hamlin, ]., Lennert, K ., Rilke, F., Stansfeld, A. G., and van Unnik, ]. A. M.: Classification of non-hodgkin's lymphoma. Lancet II, 406-408 (1974) Graumann, W.: Zur Standardisierung des SchiH'schen Reagens. Z. Wiss. Mikr. 61, 225 -226 (1953) Hale, A. ]., and Wilson, S. ].: The deoxyribonucleic acid content of leucocytes in normal and leukemic blood. ]. Path. Bact. 77, 605-614 (1959) Jones, S. E., Fuks, Z., Bull, M., Kadin, M. E., Dorfman, R. F., Kaplan, H. S., Rosenberg, S. A., and Hun Kim: Non Hodgkin's lymphomas IV. Clinicopathological correlation in 405 cases. Cancer ]I , 806-823 (1973) Knoth, W., and Sandritter, W.: Cytophotometrische Messungen des DNS-Gehaltes und morphologische Untersuchungen bei Reticulose, Reticulo-Sarcomatose und HodgkinSarkom der Haut. Arch. klin. expo Dermato!' 223, 217-235 (1965) Lennert, K., In: Leukamien und maligne Lymphome (Ed. A. Stacher), pp. 181-194. Urban & Schwarzenberg, Munchen-Berlin-Wien (1973) Lennert, K., und Mori, N.: Histologische Klassifizierung und Vorkommen des M. Hodgkin. Internist 15, 57-65 (1974) Lennert, K, Mori, N., Stein, H., and Kaiserling, E.: The histopathology of malignant lymphoma. Brit.]' Haemato!' 31 (Suppl), 193-203 (197P) Lennert, K, Stein, H., and Kaiserling, E.: Cytological and functional criteria for the classification of malignant lymphoma. Brit. ]. Cancer 31 (Suppl. II), 29-43 (1975b) Lukes, R. J., and Tindle, B. H .: Immunoblastic lymphadenopathy: A hyperimmune entity resembling Hodgkin's diesease. New Eng!. J . Med. 292, 1-8 (1975) Mathe, G., Rappaport, H., O'Conor, G. T., and Torloni, H.: Histological and cytological typing of neoplastic diseases of haematopoetic and lymphoid tissues. International Histological Classification of Tumors No. 14, World Health Organisation, Geneva (197 6) Mitrou, P. S., Queisser, W., Lennert, K., und Sandritter, W. : Kombinierte autoradiographisch-cytophotometrische Untersuchungen von Keimzentrumzellen der menschlichen Tonsillen. Virchows Arch. B 3,156- 170 (1969) Musshoff, K., Schmidt-Vollmer, S., Lennert, K., and Sandritter, W.: Preliminary clinical findings of the Kiel-classification of malignant lymphomas. Z. Krebsforsch. 87, 229 -23 8 (1976) Nathwani, B. N., Hun Kim, and Rappaport, H.: Malignant lymphoma lymphoblastic. Cancer 38, 964-983 (1976) Radaszkiewicz, T., and Lennert, K.: Lymphogranulomatosis X. Dtsch. med. Wschr. 100, 1157- 116 3 (1975) Rappaport, H., Winter, W. ]., and Hicks, E. B.: Follicular lymphoma. Are-evaluation of its position in the scheme of malignant lymphoma, based on a survey of 253 cases. Cancer 9, 792-821 (195 6) Wilms, H.: Beitrag zum Germinoblaston (Nodulares Lymphom, Brill-Symmers). Cytophotometrische Untersuchungen. Inauguraldissertation, Kiel (1970) Received January 15, 1977 . Accepted March 2, 1977
Key words: Non-Hodgkin-Lymphoma - DNA - Cytophotometry - Classification Professor Dr. Walter Sandritter, Pathologisches Institut, Albertstr. 19, D-78oo Freiburg i. Br., Germany
Original Papers
Department of Pathology (Ludwig-Aschoff-Haus), University of Freiburg (Germany)
DNA in Non Hodgkin-Lymphoma A Cytophotometric Study W. SANDRITTER and H. GRIMM
Summary Cytophotometric DNA determinations on 26 cases of non-Hodgkin's lymphoma yielded the following findings: r. Follicular centrocyticicentroblastic lymphomas (M. Brill-Symmers) and diffuse centro-
cytic lymphomas (lymphocytic lymphosarcoma) have a diploid DNA stem line. Diploid DNA values are observed in benign tumors, so that the assignment of these lymphomas to the group of "low grade malignancies" appears justified. 2.
Lymphoblastic sarcomas show an aneuploid DNA stem line, as do 96% of all malignant tumors.
3. Lymphoid cells, plasma cells, and immunoblasts seen in immunocytomas are aneuploid. Thus these lymphomas must belong to the group of "high-grade malignant lymphomas" as regards their DNA distribution. 4. Immunoblastic sarcomas have aneuploid DNA stem lines (r case tetraploid), in which both the lymphoid cells and the plasma cells from those immunoblastic sarcomas arising from immunocytomas show atypical DNA distribution patterns. 5. In two cases of angioimmunoblastic lymphadenopathy, the lymphoid cells, plasma cells, and immunoblasts arc aneuploid. They are thus regarded as "high grade malignancy" lymphomas. The results are discussed with respect to climcal course and prognosis. Measurements on a larger series of cases and correlation to clinical data are needed to support these results. Ultrafast DNA measurements made by flow-through cytophotometry can perhaps be helpful in the future for making the decision between a "low" or "high" grade malignant lymphoma.
Significant controversy surrounds the classification of the non-Hodgkin's lymphomas (malignant lymphomas, ML). The principal classification employed in English-speaking countries is that of Rappaport et al. (1956), added to which the World Health Organization (WHO) has submitted its own suggestions (Mathe et aI., 1976), and the classification by Lennert 15 Beitr. Path. Vol. 160
2
14 . W. Sandritter and H. Grimm
(the so-called "Kiel Classification", Gerard-Marchant et aI., 1974) has been under discussion since I 974. A classification of non-Hodgkin's lymphomas should not simply satisfy academic interests, but should involve prognostic perspectives as well. The Kiel classification satisfies this requirement - it identifies low grade malignancy MLs and high grade malignancy MLs - and numerous studies have appeared showing the clinical course of cases classified according to Rappaport. The goal of the following study is to contribute to the classification and to the problem of prognosis in MLs on the basis of cytophotometric DNA measurements. With regard to classification, we would hope to decide in some cases whether we are dealing with a uniform or a mixed cell population. With regard to prognosis, we are supported by our earlier investigations (B6hm and Sandritter, 1975) which have shown that benign tumors always exhibit a diploid DNA distribution pattern whereas 96% of malignant tumors show aneuploid frequency peaks. The DNA di~tribution patterns in 26 non-Hodgkin's lymphomas were investigated with respect to these two questions.
Materials and Methods Imprints were made from formalin-fixed lymph node biopsies. Lymphocytes from cases of non-specific lymphadenitis served as reference cases for the diploid, 2C DNA value. Cases were fixed for varying lengths of time in formalin. The imprints were fixed for one hour in a solution containing methanol 85%, formalin 10%, and glacial acetic acid 5%. Hydrolysis for the Feulgen reaction was performed in 4 N HCI for 60 min at 28° C. After the usual Feulgen procedure (Grauman, 1953), the slides were mounted in Cargille refraction fluid, nD 1.54. Cytophotometric measurements were made on an integrating microdensitometer (Barr and Stroud, Glasgow). The measuring point diameter was 0.5 !Am, the numerical aperture of the condenser was 0.3, and the numerical aperture of the objective was 1.3. The resulting measurements of DNA content are given in arbitrary units (AU). Between 100 and 200 cells were measured for each case. Error: Measurements made in ten diploid liver cells resulted in a standard deviation of ± 3%. The overall error, involving fixation, staining, and measurement errors, is estimated at ± 15-20%. Deviations from the diploid lymphocyte value (2C) were only counted as "true" deviations when they lay outside this 20% range.
Results Twenty-six cases of non-Hodgkin's lymphoma were measured. Diagnostic assignments for each case were made according to the old nomen-
DNA in Non Hodgkin-Lymphoma.
215
clature, the nomenclature of Lennert (Kiel Classification, 1974), and according to the classification of Rappaport et al. (1956)1). M. Brill-Symmers (ML follicular, centrocytic centroblastic; ML nodular, poorly differentiated) 1.
In all four cases studied, we were dealing with typical, follicular lymphomas with a predominant growth of germinocytes (centrocytes). Germinoblasts were present only sparsely. Both cell types were readily distinguished in the imprint preparations, and were measured separately. Figure I shows the DNA distribution pattern of a typical case (15624/75), which is representative for all cases. We see here that the DNA value for germinocytes (centrocytes) is very close to that for lymphocytes, and shows only a limited scatter toward the right up to 26 AU (diploid 18 AU, germinocyte peak at 22 AU). The germinoblasts (centroblasts) have values in the diploid region to a lesser extent, and scatter very strongly over the tetraploid value (S-phase cells) without a distinct peak formation. Table I shows that the same distribution holds for the other cases: the germinocytes are diploid with limited scatter, while the germinoblasts exhibit both diploid and interdiploid to tetraploid values. Two cases showed a hyperdiploid (triploid) peak formation.
Table 1. M. Brill-Symmers (ML follicular, centrocytic, centroblastic; ML nodular, poorly differentiated) Nr.
Age
Sex
DNA distribution pattern
15 62 4/7 6
63
~
Germinocytes diploid. Germinoblasts diploid, with scatter exceeding tetraploid (Fig. I).
16164/75
65
((
Germinocytes diploid. Germinoblasts diploid to tetraploid.
18 457/75
72
0
Germinocytes diploid. Germinoblasts triploid.
29 12 4/75
66
<3
Germinocytes diploid to triploid. Germinoblasts triploid with scatter to tetraploid. - --~--------~
1) We are very grateful to Prof. K. Lennert and Prof. H. Rappaport for reviewing our material.
216 . W. Sand ritter and H. Grimm 42
n 32 _ 28 24 _
15624/75
....
M. Brill- Symmers
20 _ 16 12 _ 8 _
Germinocyles
.
!,.. u 0
~!4
,..
E
-J
P'"
I
116t20
30
t
40
50
AU
60
20
Germinoblasls 16 12 8 4
30
t
40
50
60
Fig. I. DNA distribution pattern for germinocytes and germinoblasts in a case of M. Brill-Symmers (I5624i 75). Lymphocytes 2C = diploid = 18 AU, first arrow. Tetraploid ,= 4C = 36 AU, second arrow. n = number of cells. AU = relative DNA content in arbitrary units.
2. Lymphocytic lymphosarcoma (ML centrocytic, ML diffuse, poorly differentiated) The four cases studied showed the typical histological picture of a lymphocytic lymphosarcoma, in which the cells exhibit the appearance of centrocytic cells with very angular nuclei and an easily folded nuclear membrane without visible cytoplasm (see Lennert et al., 1975a, Fig. 3). Figure 2 (26507/7 5) gives a typical DNA distribution pattern with a peak at 30 to 32 AU (lymphocytes 28 AU). The peak of the DNA values thus lies within the variation of the diploid value. The limited scatter to higher values can be attributed to S-phase cells. All other cases showed practically the same DNA distribution pattern, without a clear peak formation in the
DNA in Non Hodgkin-Lymphoma.
26507/75
21 7
Lymphocytic Lymphosarcoma
n
24 20 16
12
8 4
40
50
t
70
60
AU
Fig. 2. DNA histogram for a lymphocytic lymphosarcoma (26507/75). Diploid AU, first arrow. Tetraploid = 56 AU, second arrow.
28
tetraploid region. Case 26507/75 had a limited amount of sclerosis, so that the lymphocytic, diffuse form of lymphosarcoma as described by Bennett and Millett (1969) is present. The patient expired five months after the initial diagnosis was made on the basis of a cervical lymph node, with autopsy findings of a generalized lymphosarcoma: lymph nodes, stomach, kidney, retroperitoneum, heart, bone.
Table ated)
2.
Lymphocytic lymphosarcoma (ML centrocytic; ML diffuse, poorly differenti-
Nr.
Age
26 50 7/75 10 301 /75
54 68
120 51/75
57
28 94 0 /75
44
Sex
DNA distribution pattern
o
Diploid with scat
o o o
'~r
to triploid values (Fig. 2).
Diploid with scatter to triploid values. Diploid. Diploid with scatter to tetraploid values.
3. Lymphoblastic lymphosarcoma (ML lymphoblastic or undifferentiated) This group of tumors is readily distinguished in terms of its DNA distribution pattern from the lymphocytic lymphosarcomas. All four cases show peak formation at hyperdiploid or triploid values.
2 18 •
Table 3.
W. Sandritter and H. Grimm Lymphoblastic lymphosarcoma (ML lymphoblastic)
Nr.
Diagnosis
Age
Sex
DNA distribution pattern 1)
353 8/7 6
Burkitt
I2
0
Hyperdiploid. Peak 38 AU.
24 I 53/75
Lymphoblastic
I5
~
Hyperdiploid (Peak at 36 AU), second peak around 70 AU (Fig. 3).
1735?/75
Lymphoblastic
60
0
Hyperdiploid (Peak at 38 AU), second peak around 70 AU.
897/75
Lymphoblastic
68
~
Triploid (43 A), second peak around 84 AU (Fig. 4).
1) Diploid value 28 AU, tetraploid 56 AU, octoploid II2 AU.
Case 3538/76 involves a histologically typical Burkitt's lymphoma with a frequency peak of DNA values at 38 AU hyperdiploid and a wide scatter of values up to 40-60 without a second peak (Tab. 3). The remaining three cases (24153/75,897175,17357175) are unclassifiable lymphoblastic sarcomas (non-convoluted; after Rappaport: diffuse, lymphoblastic, undifferentiated). The DNA distribution for case 24153/75 shows a hyperdiploid peak (36 AU) with wide scatter up to an implicit doubling-value of
n 40
24153/75
Lymphoblastic
Lymphosarcoma
36 32 28 24 20 16 12 8 4
28
t
30
40
50
t
60
70
80
90
AU
Fig. 3. Lymphoblastic lymphosarcoma (24153/75) with a hyperdiploid (36 AU) DNA stem line (First arrow diploid value; second arrow tetraploid).
DNA in Non Hodgkin-Lymphoma. 219
n
897/75 Lymphobl..tlc Lymphosarcoma
40 36 32
28 24 20 16
12
8
100
120
AU
Fig. 4. Lymphoblastic lymphosarcoma with a triploid DNA stem line (42 AU) and a second peak at circa 80 AU (hypooctoploid). First arrow = tetraploid value. Second arrow = octoploid value.
70 AU (hypertetraploid) (see Fig. 3). Case 17357/75 shows the same pattern (peak at 38 AU with a weak doubling peak), and in case 897/75 (Fig. 4) we find a triploid peak (42 AU with doubling at 80 AU). Cells in GI phase (hyperdiploid or triploid) were distinguishable from cells in G2 phase (in the doubling peak) in all three cases by means of a larger nucleus and more open chromatin pattern.
4. Immunocytoma (ML lympho-plasmocytic-plasmacytoid) Both we and Lennert made the diagnosis of lympho-plasmacytoid immunocytoma for these four cases (829/75 and 2929I!75 polymorphic type). Rappaport made varying diagnoses (see Tab. 4), where it should be recognized that methyl-green pyronine slides were not available to Rappaport. All cases showed the typical, histological picture: diffuse infiltration with lymphoid cells which were not further classifiable, small lymphoid plasma cells (typical as well as large, atypical plasma cells), and large immunoblasts with basophilic cytoplasm and 1-2 large nucleoli (see Lennert et al., 1975b, Fig. 9). The various cell types were distinguishable only to a limited extent on the basis of nuclear morphology in the Feulgen-stained smear preparations. Our cytophotometric measurements were thus restricted to distinguishing between lymphoid cells and small plasma cells on the one hand and cells with large nuclei (large plasma cells and immunoblasts) on the other.
130
220 . W. Sandritter and H. Grimm Table 4.
lmmunocytoma (ML Iympho-plasocytic-plasmocytoid; ML !'p.)
~-------~-,------.
No.
Diagnosis
Age
851 0 /7 6
ML!.p
54
769/75
ML !.p2)
82 9/75
ML !'p3)
29 2 9
r175
Sex
DNA-distribution pattern l ) Hyperdiploid 38-40 AU and hypertetraploid (around 60 AU) (Fig. 5). Hyperdiploid 38 AU and hypertetraploid 60-70 AU (Fig. 6).
54
ML !.p4)
Hypotetraploid 44 AU. Hypooctoploid 86 AU. Hyperdiploid 34 AU scattered values at 76 AU.
- _ . _.._----- -- ----------
1) 2} 3) 4)
Diploid Value 28 AU, tetraploid 56 AU. After Rappaport: ML diffuse, well differentiated, lymphocytic. After Rappaport: ML diffuse, mixed. After Rappaport: ML diffuse, well differentiated, lymphocytic. 4) After Lennert: ML Iympho-plasmocytoid, polymorphic. ad 2) 769/75 Autopsy findings: tumor infiltrates in all lymph nodes, skin, stomach, gut, liver, kidneys, spleen, bone marrow. Course: 6 years. ad 3) 829/75 Autopsy findings: para aortic lamph nodes, peritoneum. Course: 6 months. Polymorphic type.
Table 4 shows that the DNA distribution pattern is rather similar for each of the four cases (see Figs 5 and 6). The DNA values for the small lymphoid cells and lymphoid plasma cells show their first frequency peak in the hyperdiploid region (hypotetraploid for case 829/75), while the large nucleus cells (atypical or typical plasma cells, immunoblasts) exhibit a second frequency peak at 60-70 AU (hypertetraploid).
5. I mmunoblastic sarcomas This group of tumors has had a variety of designations: reticulum cell sarcoma (Mathe et aI., I976), ML diffuse histiocytic (Rappaport et aI., I956), and immunoblastic sarcoma (Lennert, Kiel Classification). Our 8 cases consist in part of pure immunoblastic sarcomas, as read by Rappaport, which, however, were diagnosed as malignant lymphomas of the centroblastic type by Lennert (see Tab. 5). Four cases involved immunoblastic sarcomas arising from an immunocytoma, so that they are rich in lymphoid cells, plasma cells, and small and large immunoblasts. Figure 7 gives an example of the DNA distribution pattern of the immunoblastic sarcomas arising from an immunocytoma (23395/75). The lymphoid cell elements and the small immunoblasts show widely scattered
DNA in Non Hodgkin-Lymphoma·
221
n 40 36
8510/76
Immunocytoma
32 28 24 20 16 12 8 4
t
20
30
t
50
40
60
70
AU
Fig. 5. Immunocytoma (8510/75) with a hyperdiploid DNA stem line (38-40 AU) (lymphoid cells) and a second peak in the hypertetraploid region (60-70 AU). First arrow = diploid. Second arrow = tetraploid.
n
769/75
Immunocytoma
40 36 32 28
20 16
12
30
40
50
t
60
70
80
90
100
120
AU
Fig. 6. Immunocytoma (769/75) with a hyperdiploid DNA stem line (38 AU) (lymphoid cells) and hypertetraploid DNA values (60-70 AU). Wide scatter of values up to 130 AU (112 AU = octoploid). First arrow = tetraploid value. Second arrow = octoploid value.
130
222 .
W. Sandritter and H . Grimm
n 40
23395/75
Immunoblast ic Sarcoma
36 32
28 24 20 16 12 8 4
t
30
40
50
100
AU
Fig. 7. Immunoblastic sarcoma arising from an immunocytoma (23395 /75). Lymphoid cells and small immunoblasts show values in the triploid to hypotetraploid region 40-50 AU). Second peak at 80 AU. First arrow = diploid value. Second arrow = tetraploid value.
n
28091/7'
Immunoblestic Sarcoma
32 28 24 20 1&
12 8
30
40
50
t
eo
70
80
90
100
Fig. 8. Immunoblastic sarcoma (2609I /75) with a tetraploid DNA stem line (56 AU) and wide scatter up into the octoploid region (I I2 AU). First arrow = tetraploid value. Second arrow = octoploid value.
DNA in Non Hodgkin-Lymphoma. 223
values (40-50 AU = triploid-hypotetraploid), and a doubling peak is seen at 80 AU (Fig. 7). All other cases show practically the same DNA distribution pattern, with few variations (see Tab. 5). Figures 8 and 9 shows immunoblastic sarcomas without lymphoid cells, having only immunoblasts (small immunoblasts in GI phase and large immunoblasts in G2 phase) with basophilic cytoplasm and centrally positioned nucleoli in a clear nucleus. Case 2609ri75 (Fig. 8) is unambiguously tetraploid with a weak doubling peak at IOO-I2O AU. Case 18082/75 (Fig. 9) shows triploid values with wide scatter and no second peak formation. The two remaining cases behave in a similar fashion.
Table 5. No.
Immunoblastic Sarcomas (IS) (Reticulosarcoma, ML diffuse, histiocytic) Diagnosis
Age
Sex
DNA Distribution pattern 1)
- - - - - - ---- - - - -
IS arising from an immunocytoma (spleen, liver)
67
Lymphoid cells and small immunoblasts with triploid to hypotetraploid values (40-50 AU). Doubling peak at 80 AU (Fig. 7).
IS from immunocytoma
75
Lymphoid cells with triploid values. Small immunoblasts with hypotetraploid values. No clear, second peak. Values from 60 to 132 AU.
IS from immunocytoma
7I
Hyperdiploid (36 AU). No clear second peak. Scatter to 76 AU.
IS from immunocytoma (stomach)
73
Lymphoid cells and small immunoblasts with hyperdiploid values (36 AU). Immunoblasts peak at 72 AU, wide scatter.
IS2)
71
Tetraploid (56 AU). Values at 100-130 AU (Fig. 8).
18082/75
IS3)
35
206 38/75
IS4)
53
o o
Triploid (42 AU). Marked scatter (Fig.9). Tripoloid (42 AU). Values between 70 and 82 AU. Hyperdiploid (36 AU). Wide scatter, no second peak.
---~----~~---
1) Diploid Values 28 AU, tetraploid Values 56 AU. 2) Our diagnosis: immunoblastic sarcoma. Rappaport: ML diffuse histiocytic. Lennert: ML centroblastic. 3) Same diagnosis as 2. -1) Same diagnosis as 2.
224 . W. Sand ritter and H. Grimm n 32 18082/75
Immunoblastlc Sarcoma
28 24 20 18 12
8 4
t
~o
70
80
80
90
100
AU
Fig. 9. Immunoblastic sarcoma (18082/ 75) with a triploid DNA stem line and wide scatter of DNA values. First arrow = diploid value. Second arrow = tetraploid value.
n 18977/75 36
Angiolmmunoblastlc Lymphadenopathy
32 28 24
20 18 12
8 4
1
30
40
50
i
60
70
80
90
AU
Fig. 10. Angioimmunoblastic lymphadenopathy (18977/ 75). Lymphoid cells and plasma cells are near-triploid, immunoblasts are hypertetraploid. First arrow = diploid value Second arrow = tetraploid value.
6. Angioimmunoblastic lymphadenopathy (Lymphogranulomatosis X) This lymphoma, described by Frizzera et al. (I974) and Lukes et al. (I975), had been designated as lymphogranulomatosis X by Lennert in
DNA in Non Hodgkin-Lymphoma .
225
n 40 36 32 28
19095/7S
24
Angioimmunoblastic
lymphadenopathy 20
lymphoid cells
16 12
8 4
30
40
so
30
40
50
t
AU
60
24 20 16 12 8
t
60
70
80
90
AU
Fig. I I. Angioimmunoblastic lymphadenopathy (19095 /75). Lymphoid cells are hyperd iploid. Immunoblasts are triploid (42 AU) . 28 AU = diploid value. Arrow = tetraploid value.
1973. Our two cases show a proliferation of venules and a mixed cellular infiltrate of lymphoid cells, plasma cells, immunoblasts, histiocytes, and giant cells, which can resemble Sternberg giant cells. The Feulgen preparations only allow us to distinguish lymphoid cells and plasma cells on the one hand from small and large immunoblasts on the other, where the latter are characterized by their looser chromatin pattern and a central nucleolus. The lymphoid cells and small immunoblasts show a near-triploid peak in case 18977/75, while the larger immunoblasts scatter stongly to higher values up to 70-80 AU (Fig. 10). The limited scatter in the small lymphoid cells is particularly noteworthy. Case 19095/75 shows a similar DNA distribution pattern (Fig. r r).
226 . W. Sand ritter and H. Grimm Table 6.
Angioimmunoblastic lymphadenopathy (AL) (Lymphogranulomatosis X) ~-~~~---~---
No.
Diagnosis
~----~--
Age
~-~----------
Sex
--~-~~~~-
DNA-Distribution Pattern Lymphoid cells, plasma cells with near triploid values (40 AU). Immunoblasts hypertetraploid (70-80 AU). Wide scatter (Fig. r 0).
79
AL2)
Lymphoid cells hyperdiploid (38 AU). Immunoblasts 40-42 AU. Wide scatter. (Fig. r r). -----~
-------~
- - -
1) After Lennert: Al or T-Zone Lymphoma. 2) Rich in Plasma cells.
Discussion Almost no studies have been performed on the DNA-distribution pattern of the non-Hodgkin's lymphomas. The above cytophotometric measurements made on 26 cases sought to answer the extent to which it is possible to judge the malignancy of these lymphomas on the basis of the DNA distribution pattern (euploidy, aneuploidy). We further sought to answer whether a given cell population was uniform or mixed. It is necessary to realize that for DNA values lying between diploid (2C) and tetraploid (4C), or between tetraploid and octoploid (8c), we cannot decide whether we are dealing with cells in normal S-phase as opposed to aneuploid cells. Furthermore, DNA content makes no predictions about chromosomal abnormalities. Since we had available only formalin fixed material instead of fresh imprints, we were dealing with an error of about the diploid value of 15-20%. Finally, it is known from numerous prior investigations that lymphocytes often give 10% lower DNA values thaa, say, diploid liver cells (problem of chromatin condensation). Follicular lymphomas, centroblasticlcentrocytic (M. Brill-Symmers) showed two cell populations. Centrocytes were diploid (within the 15 to 20% range of scatter), while the centroblasts (germinoblasts) scattered up to tetraploid values. Wilms (1970) found the same distribution pattern in four cases of Brill-Symmers which were studied in Lennert's department. The interdiploid-tetraploid DNA values seen in centroblasts are presumably attributable to S-phase cells. Mitrou et al. (1969) found that normal germinoblasts from normal germinal centers showed 3H thymidine uptake in up to 50% of the nuclei, and the corresponding DNA contents for these cells lay in the interdiploid-tetraploid range. The corresponding germino-
DNA in Non Hodgkin-Lymphoma.
227
cytes showed only a very limited 3H thymidine uptake (8%). In this regard, we consider the rather indistinct peak formation of DNA values in the triploid range seen in two cases not as an aneuploid DAN stem line, but more likely a random accumulation of S-phase values (partial synchroni. .. ~) zatlon We conclude from the four cases we studied and from the measurements made by Wilms that these cases are characterized by an euploid distribution pattern, compatible with the relatively benign character of these lymphomas (low malignancy, according to the Kiel classification). Previous studies in our laboratory (Bohm and Sandritter, 1975) have shown that benign tumors are always euploid. Jones et al. (1973) also found a significantly higher survival rate in nodular lymphomas as compared to the diffuse lymphomas. Further studies on a larger number of cases will be necessary in order to reach a definitive conclusion on this matter 1). We found an entirely similar DNA distribution pattern for lymphocytic lymphosarcomas (ML diffuse centrocytic). All cases were diploid with only a limited scatter of DNA values (d. Hale et al., 1959). These entities are also rated as low-grade MLs by the Kiel Classification. Jones et al. (1973) found a lower survival rate for the diffuse malignant lymphomas. Musshoff et al. (1976) were unable to show a difference in the acturial survival rate for nodular and diffuse centrocytic malignant lymphomas. Lymphoblastic lymphosarcomas belong to the high grade MLs. Nathwani et al. (1976) found a mean survival rate of 8 months. All four cases in this study (one Burkitt's and three non-convoluted) showed an aneuploid DNA distribution. (DNA stem line in the hyperdiploid or triploid region, often with a hypertetraploid doubling peak). We found such aneuploid DNA stemlines only in malignant tumors in our ealier studies (Bohm and Sand ritter, I975). In other words, the DNA measurements corroborate the high grade malignancy of these MLs. The immunocytomas exhibit a mixed cell population, for which it is difficult to distinguish the individual components in the Feulgen-stained preparations. Our DNA measurements show that the immunocytomas consist of aneuploid cells as do other malignant tumors (Bohm and Sandritter, 1975), containing plasma cell, lymphoid cell, and immunoblastic components. The extent to which the cells in the first and second frequency peaks 1) The results of our observations are only partially in agreement with the DNA determinations of R. Silvestrini et al. (J. Nat. Cancer Inst. 58, 499, 1977) on nonHodgkin's lymphoma. In M. L. immunoblastic, lymphoblastic and immunocytoma, as well as in M. L. centrocyticlcentroblastic, these workers found diploid DNA values. At the moment we are unable to explain this discrepancy.
228 .
W. Sandritter and H. Grimm
(hyperdiploid, hypertetraploid) arise at least in part from a common cell population cannot be decided from the measurements. We suspect that small immunoblasts are hidden among the small lymphoid cells, and that these immunoblasts form the G2 cells in the second peak, but this second peak could equally well consist of plasma cells. Immunocytomas are regarded as low grade malignancy MLs after the Kiel Classification. We have two cases with good follow-up information: one case (769/75) survived 6 years, the other (829/ 75) survived 6 Months. Musshoff et al. (1976) found a better survival rate for diffuse lymphoplasmocytoid immunocytomas, but this is not statistically different from that of the high grade malignant lymphomas. The DNA measurements (aneuploid DNA distribution pattern with DNA stem lines) assign the immunocytomas to the group of high grade malignant lymphomas. Further studies will be necessary before reaching a final judgment on this matter. The transformation to an immunoblastic sarcoma is of particular interest, especially arising from a polymorphic immunocytoma. The immunoblastic sarcomas are morphologically a heterogeneous group. Immunoblastic sarcomas which arise from immunocytomas show typically hyperdiploid or triploid lymphoid cells (lympoid cell elements, plasma cells) and small and large immunoblasts which may have triploid or hypotetrapolid DNA-values with a doubling peak (hypooctoploid). We are obviously dealing with two aneuploid cell populations. Pure immunoblastic sarcomas may show hyperdiploid, triploid, or tetraploid DNA stem lines. Lennert diagnosed a centroblastic ML in three of four cases. The cytophotometric measurements make no distinction between immunoblastic sarcoma and centroblastic ML. The immunoblastic sarcomas were earlier regarded as reticulum cell sarcomas. We cannot be certain that true reticulum cell sarcomas do not arise in lymph nodes. Reticulum cell sarcomas of the skin seem to be exceptional in their clinical behavior as well (protracted course). Our own DNA measurements (Knoth et al., 1965) made on a reticulum cell sarcomatosis in the skin showed a tetraploid DNA stem line, as in the above case 26091 / 75· Two cases of angioimmunoblastic lymphoadenopathy (AL) were found to be cytophotometrically heterogeneous as well. Lymphoid cells and plasma cells showed a hyperdiploid or triploid DNA stem line, while the immunoblasts showed a wide scatter of values. Thus AL showed a DNA distribution pattern characteristic of a malignant tumor. Clinically, AL appears to be a highly malignant lymphoma, with a course of 1-18
DNA in Non Hodgkin-Lymphoma . 229
months (Lennert et al., 1974) or 1-45 months (Radaszkiewicz and Lennert, 1975). Further studies on a larger series of MLs must be done in order to show the extent to which cytophotometric DNA measurements, correlated with observations of clinical course, can offer a more accurate prognostic judgment. Lymph node material can also readily be worked up for DNA measurements by means of flow-through cytophotometers (pulse cytophotometry), so that a rapid evaluation with immediate diagnostic feedback becomes possible.
Zusammenfassung Zytophotometrische DNS-Bestimmungen an 26 Fallen von Non-Hodgkin-Lymphomen ergaben: 1.
Follikulare zentrozytischlzentroblastische Lymphome (M. Brill-Symmers) haben ebenso wie die diffusen zentrozytischen Lymphome (lymphozytisches Lymphosarkom) eine diploide DNS-Stammlinie. Diploide DNS-Werte werden in gutartigen Tumoren beobachtet, so dag die Einstufung dieser Lymphome in der Gruppe "low grade malignancy" gerechtfertigt erscheint.
2.
Lymphoblastische Sarkome zeigen eine aneuploide DNS-Stammlinie, wie 96 % malignen Tumorcn.
aller
3. Bei den Immunozytomen sind lymphoide Zellen und Plasmazellen ebenso wie die Immunoblasten aneuploid. Die Lymphome miifhen, nach der DNS-Verteilun~ zu urteilen, zu den "High-grade-malignancy" -Lymphomen zu rechnen sein. 4. Immunoblastische Sarkome zeigen aneuploide DNS-Stammlinien (I Fall tetraploid), wobei auch die lymphoiden Zellen und Plasmazellen bei den immunoblastischen Sarkomen, die aus Immunozytomen entstanden sind, atypische DNS-Verreilungsmuster aufweisen. 5. In zwei Fallen von angioimmunoblastischer Lymphadenopathie sind lymphoide Zellen, Plasmazellen und Immunoblasten aneuploid. Sie miissen demnach als "High-grademalignancy"-Lymphome angesehen werden. Die Ergebnisse werden im Zusammenhang mit dem klinischen Verlauf und der Prognose diskutiert. Messungen an einer grogeren Zahl von Fallen und Korrelation zu den klinischen Daten solhen diese Ergebnisse untermauern. Uhraschnelle DNS-Messungen mit der Durchflugphotometrie konnen vielleicht in Zukunft hilfreich sein fur die Entscheidung "low" oder "high malignancy" lymphom.
Literatur Bennett, M. H., and Millett, Y. L.: Nodular sclerotic lymphosarcoma. A possible new clinico-pathological entity. Clin. Radio!. 20, 339-343 (1969) Bohm, N., and Sandritter, W.: DNA in human tumors: A cytophotometric study. Curro Top. Path. 60, 152-214 (1975) (Springer-Verlag, Heidelberg) Frizzera, G., Moran, E. M., and Rappaport, H.: Angio-immunoblastic lymphadenopathy with dysproteinaemia. Lancet 1070-1073 (1974) 16 Beitr. Path. Vol. 160
230 . W. Sandritter and H. Grimm Gerard-Marchant, R., Hamlin, ]., Lennert, K ., Rilke, F., Stansfeld, A. G., and van Unnik, ]. A. M.: Classification of non-hodgkin's lymphoma. Lancet II, 406-408 (1974) Graumann, W.: Zur Standardisierung des SchiH'schen Reagens. Z. Wiss. Mikr. 61, 225 -226 (1953) Hale, A. ]., and Wilson, S. ].: The deoxyribonucleic acid content of leucocytes in normal and leukemic blood. ]. Path. Bact. 77, 605-614 (1959) Jones, S. E., Fuks, Z., Bull, M., Kadin, M. E., Dorfman, R. F., Kaplan, H. S., Rosenberg, S. A., and Hun Kim: Non Hodgkin's lymphomas IV. Clinicopathological correlation in 405 cases. Cancer ]I , 806-823 (1973) Knoth, W., and Sandritter, W.: Cytophotometrische Messungen des DNS-Gehaltes und morphologische Untersuchungen bei Reticulose, Reticulo-Sarcomatose und HodgkinSarkom der Haut. Arch. klin. expo Dermato!' 223, 217-235 (1965) Lennert, K., In: Leukamien und maligne Lymphome (Ed. A. Stacher), pp. 181-194. Urban & Schwarzenberg, Munchen-Berlin-Wien (1973) Lennert, K., und Mori, N.: Histologische Klassifizierung und Vorkommen des M. Hodgkin. Internist 15, 57-65 (1974) Lennert, K, Mori, N., Stein, H., and Kaiserling, E.: The histopathology of malignant lymphoma. Brit.]' Haemato!' 31 (Suppl), 193-203 (197P) Lennert, K, Stein, H., and Kaiserling, E.: Cytological and functional criteria for the classification of malignant lymphoma. Brit. ]. Cancer 31 (Suppl. II), 29-43 (1975b) Lukes, R. J., and Tindle, B. H .: Immunoblastic lymphadenopathy: A hyperimmune entity resembling Hodgkin's diesease. New Eng!. J . Med. 292, 1-8 (1975) Mathe, G., Rappaport, H., O'Conor, G. T., and Torloni, H.: Histological and cytological typing of neoplastic diseases of haematopoetic and lymphoid tissues. International Histological Classification of Tumors No. 14, World Health Organisation, Geneva (197 6) Mitrou, P. S., Queisser, W., Lennert, K., und Sandritter, W. : Kombinierte autoradiographisch-cytophotometrische Untersuchungen von Keimzentrumzellen der menschlichen Tonsillen. Virchows Arch. B 3,156- 170 (1969) Musshoff, K., Schmidt-Vollmer, S., Lennert, K., and Sandritter, W.: Preliminary clinical findings of the Kiel-classification of malignant lymphomas. Z. Krebsforsch. 87, 229 -23 8 (1976) Nathwani, B. N., Hun Kim, and Rappaport, H.: Malignant lymphoma lymphoblastic. Cancer 38, 964-983 (1976) Radaszkiewicz, T., and Lennert, K.: Lymphogranulomatosis X. Dtsch. med. Wschr. 100, 1157- 116 3 (1975) Rappaport, H., Winter, W. ]., and Hicks, E. B.: Follicular lymphoma. Are-evaluation of its position in the scheme of malignant lymphoma, based on a survey of 253 cases. Cancer 9, 792-821 (195 6) Wilms, H.: Beitrag zum Germinoblaston (Nodulares Lymphom, Brill-Symmers). Cytophotometrische Untersuchungen. Inauguraldissertation, Kiel (1970) Received January 15, 1977 . Accepted March 2, 1977
Key words: Non-Hodgkin-Lymphoma - DNA - Cytophotometry - Classification Professor Dr. Walter Sandritter, Pathologisches Institut, Albertstr. 19, D-78oo Freiburg i. Br., Germany