J Clin Pathol 1993;46:915-921 Haemopoietic regrowth after chemotherapy for acute leukaemia: an immunohistochemical study of bone marrow trephine biopsy specimens B S Wilkins, A G Bostanci, M F Ryan, D B Jones Abstract Aim-To analyse haemopoietic regrowth and residual disease in bone marrow trephine biopsy specimens after treat- ment for acute leukaemia, using immunohistochemical staining. Methods-Biopsy specimens before and after treatment were studied from patients diagnosed as having acute myeloid or lymphoblastic leukaemia. Specimens after treatment encompassed periods from two to 56 weeks from the start of treatment. Routine haematoxylin and eosin and Giemsa stained sections were evaluated in association with immunostained preparations. A panel of antibodies was used, which reacts with epitopes showing restricted expression dependent on the lineage or maturation stage of cells. Results were evaluated in the light of clinical, peripheral blood, and marrow aspirate findings. Results-The speed and sequence of regrowth of haemopoietic cells were more variable than expected. Immuno- staining highlighted features of dysplasia after treatment and in some cases assisted detection of residual or relapsed leukaemia. Peripheral blood and aspirate cell counts reflected accurately the amount of regrowth, but not the dyspla- sia, seen in biopsy samples. Delayed regrowth was associated with complex individual factors. Conclusions-Morphological and immuno- histochemical study of trephine biopsy specimens from patients treated for acute leukaemia provides information complementary to that obtained from peripheral blood and aspirated marrow. Variation in the timing and sequence of regrowth is highlighted. Immunostaining can aid in the detection of relapse or minimal residual leukaemia. The clinical relevance of dysplastic changes in biopsy specimens after treatment is uncertain, but such changes may persist for long periods. (7 Clin Pathol 1993;46:915-921) Standard chemotherapeutic agents used to treat acute leukaemia depress normal haemopoiesis. Individual patients have cytopenias of varying severity for different lengths of time after treatment. Little is known about factors which influence the rates of regrowth of different haemopoietic cell lineages. Age, episodes of sepsis, pre-existing myelodysplasia and the nature of the chemotherapeutic agents probably all play a part, in addition to factors specific to individ- ual patients. Haemopoietic regrowth after chemother- apy is usually monitored by sequential exami- nations of peripheral blood samples. Few studies have investigated the underlying appearances in bone marrow trephine biopsy specimens,12 and little information is avail- able on the sequential changes in the marrow as regrowth occurs. As peripheral blood find- ings frequently include dysplastic cellular changes which may persist for long periods, it is of interest to know whether these reflect haemopoietic stem cell or stromal cell dam- age sustained as a result of chemotherapy. Abnormalities in the cytology or localisation of regenerating haemopoietic cells within bone marrow could provide clues to the ori- gin of such dysplastic changes and might pro- vide useful variables by which to gauge the adequacy of haemopoietic reconstitution. We and others have previously reported the use of immunohistochemical staining to demonstrate the lineage and maturation stage of haemopoietic cells within bone marrow biopsy specimens.34 In this study we used immunohistochemistry to study bone marrow trephine biopsy specimens obtained before and after treatment for acute myeloid and lymphoblastic leukaemias to assess haemo- poietic regrowth and residual disease. We have correlated our morphological findings with variables reflecting patients' clinical progress, including marrow aspirate and peripheral blood appearances, episodes of sepsis, and exposure to successive courses of chemotherapy. For most patients, at least two post-treatment biopsy specimens were available for study, and multiple biopsy specimens from some patients encompassed prolonged periods, including several episodes of disease remission and relapse. Methods Twenty patients, treated for acute myeloid or lymphoblastic leukaemia at Southampton University hospitals, were selected for study on the basis of availability of appropriate archival trephine biopsy material. Ages ranged from 2 to 71 years (mean 39 5 years). There were three children (ages 2, 2, and 10 years), two of whom were diagnosed as having acute lymphoblastic leukaemia (ALL) Department of Pathology, Southampton General Hospital B S Wilkins D B Jones Department of Haematology M F Ryan Department of Pathology, University of Ankara, Turkey A G Bostanci Correspondence to: Dr B S Wilkins, Department of Pathology, Southampton General Hospital, Tremona Road, Southampton, S09 4XY. Accepted for publication 22 April 1993 915 on August 5, 2021 by guest. Protected by copyright. http://jcp.bmj.com/ J Clin Pathol: first published as 10.1136/jcp.46.10.915 on 1 October 1993. Downloaded from
Transcript
J Clin Pathol 1993;46:915-921
Haemopoietic regrowth after chemotherapyfor acute leukaemia: an immunohistochemicalstudy of bone marrow trephine biopsy specimens
B S Wilkins, A G Bostanci, M F Ryan, D B Jones
AbstractAim-To analyse haemopoietic regrowthand residual disease in bone marrowtrephine biopsy specimens after treat-ment for acute leukaemia, usingimmunohistochemical staining.Methods-Biopsy specimens before andafter treatment were studied frompatients diagnosed as having acutemyeloid or lymphoblastic leukaemia.Specimens after treatment encompassedperiods from two to 56 weeks from thestart of treatment. Routine haematoxylinand eosin and Giemsa stained sectionswere evaluated in association withimmunostained preparations. A panel ofantibodies was used, which reacts withepitopes showing restricted expressiondependent on the lineage or maturationstage of cells. Results were evaluated inthe light of clinical, peripheral blood,and marrow aspirate findings.Results-The speed and sequence ofregrowth of haemopoietic cells weremore variable than expected. Immuno-staining highlighted features of dysplasiaafter treatment and in some casesassisted detection of residual or relapsedleukaemia. Peripheral blood and aspiratecell counts reflected accurately theamount of regrowth, but not the dyspla-sia, seen in biopsy samples. Delayedregrowth was associated with complexindividual factors.Conclusions-Morphological and immuno-histochemical study of trephine biopsyspecimens from patients treated foracute leukaemia provides informationcomplementary to that obtained fromperipheral blood and aspirated marrow.Variation in the timing and sequence ofregrowth is highlighted. Immunostainingcan aid in the detection of relapse orminimal residual leukaemia. The clinicalrelevance of dysplastic changes in biopsyspecimens after treatment is uncertain,but such changes may persist for longperiods.
(7 Clin Pathol 1993;46:915-921)
Standard chemotherapeutic agents used totreat acute leukaemia depress normalhaemopoiesis. Individual patients havecytopenias of varying severity for differentlengths of time after treatment. Little isknown about factors which influence the rates
of regrowth of different haemopoietic celllineages. Age, episodes of sepsis, pre-existingmyelodysplasia and the nature of thechemotherapeutic agents probably all play apart, in addition to factors specific to individ-ual patients.
Haemopoietic regrowth after chemother-apy is usually monitored by sequential exami-nations of peripheral blood samples. Fewstudies have investigated the underlyingappearances in bone marrow trephine biopsyspecimens,12 and little information is avail-able on the sequential changes in the marrowas regrowth occurs. As peripheral blood find-ings frequently include dysplastic cellularchanges which may persist for long periods, itis of interest to know whether these reflecthaemopoietic stem cell or stromal cell dam-age sustained as a result of chemotherapy.Abnormalities in the cytology or localisationof regenerating haemopoietic cells withinbone marrow could provide clues to the ori-gin of such dysplastic changes and might pro-vide useful variables by which to gauge theadequacy of haemopoietic reconstitution.We and others have previously reported
the use of immunohistochemical staining todemonstrate the lineage and maturation stageof haemopoietic cells within bone marrowbiopsy specimens.34 In this study we usedimmunohistochemistry to study bone marrowtrephine biopsy specimens obtained beforeand after treatment for acute myeloid andlymphoblastic leukaemias to assess haemo-poietic regrowth and residual disease.We have correlated our morphological
findings with variables reflecting patients'clinical progress, including marrow aspirateand peripheral blood appearances, episodesof sepsis, and exposure to successive coursesof chemotherapy. For most patients, at leasttwo post-treatment biopsy specimens wereavailable for study, and multiple biopsyspecimens from some patients encompassedprolonged periods, including several episodesof disease remission and relapse.
MethodsTwenty patients, treated for acute myeloid orlymphoblastic leukaemia at SouthamptonUniversity hospitals, were selected for studyon the basis of availability of appropriatearchival trephine biopsy material. Agesranged from 2 to 71 years (mean 39 5 years).There were three children (ages 2, 2, and 10years), two of whom were diagnosed ashaving acute lymphoblastic leukaemia (ALL)
Department ofPathology,Southampton GeneralHospitalB S WilkinsD B JonesDepartment ofHaematologyM F RyanDepartment ofPathology, UniversityofAnkara, TurkeyA G BostanciCorrespondence to:Dr B S Wilkins, Departmentof Pathology, SouthamptonGeneral Hospital, TremonaRoad, Southampton, S094XY.Accepted for publication22 April 1993
915
on August 5, 2021 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.46.10.915 on 1 October 1993. D
and one acute monoblastic leukaemia (AML-M5). Of the adult patients, 12 had acutemyeloid leukaemia (AML) of various sub-types. Four had ALL, and one acute undiffer-entiated leukaemia (AUL). Three of theadults had myelodysplasia preceding thedevelopment of AML. Follow up rangedfrom two to 56 weeks after diagnosis.Chemotherapy consisted of standard regi-mens in current use for AML and ALL. Twopatients underwent autologous bone marrowtransplantation during the follow up period.Two patients were retreated after relapses at17 and 21 weeks, respectively. They hadachieved remission and completedchemotherapy for the first episode ofleukaemia. Patient records were reviewed toobtain details of precise nature and timing ofchemotherapy; intercurrent illnesses, particu-larly febrile episodes and confirmed infec-tions; timing and findings of peripheral bloodcell counts, blood films, and aspirated bonemarrow; enzyme cytochemical phenotypingof leukaemic blast cells at time of diagnosis(in blood or aspirated marrow).
Seventy one biopsy specimens were suit-able for study, of which 24 were taken beforeand 47 after treatment. The number of speci-mens taken after treatment for each patientranged from one to four (most patients hadtwo). For two patients, additional biopsyspecimens were available after retreatment forrelapse, as described above.
Biopsy specimens were processed routinelyby fixation in 10% neutral buffered formalin,decalcification using 5% formic acid in for-malin, and embedding in paraffin wax. Forgeneral assessment, sections were stainedwith haematoxylin and eosin. Giemsa, and asilver stain for reticulin.
Sections 5 pm thick were mounted onslides coated with poly-L-lysine. After dryingovernight at 37°C, they were immunostainedusing a panel of antibodies selected for theirreactivities with myeloid cells of different lin-eages and maturation stages. Details of theantibodies, their sources, and major reactivi-ties are given in table 1. A well establishedstreptavidin-biotin complex immunoperoxi-dase technique was used.'3 Endogenous per-oxidase activity was blocked using methanolic
Table 1 Antibodies usedfor immunostaining, with details of their sources and specificitiesAntibody Specificity Reference Source Reactivity in marrow
LCA CD45 5 Dako Ltd Lymphoid andmyeloid cells
KP1 CD68 6 D Mason, Oxford PM andM + +Mono + +
NP57 Neutrophil 7 D Mason, Oxford PM andM + + +elastase MM and NP +
Mac387 Calgranulin 9 D Jones, Southampton MM + +NP + + +Mono + + to +++
BRIC 101 a-sialoglyco- 10 D Anstee, Bristol PE and EN +protein (a-SGP) MN, LN and RC + + +
Ret40f f-sialoglyco- 11 D Mason, Oxford PE, N and RC + + +protein (8-SGP)
Y2/51 CD61 11 DMason, Oxford MK- to +++AA1 Mast cell 12 A Walls, Southampton Mast cells
tryptase
CD = cluster of differentiation; PM = promyelocyte; M = myelocyte; MM = metamyelocyte;NP = neutrophil; Mono = monocyte; PE = pro-erythroblast; EN = early normoblast; MN =
mid-normoblast; LN = late normoblast; N = normoblast, all stages; RC = mature red bloodcell; MK = megakaryocyte.
hydrogen peroxide; trypsin digestion was per-formed where necessary for antigen demon-stration and the final coloured product(brown) was developed using 3,3' diamino-benzidine hydrochloride (DAB) as the chro-magen. Sections were lightly counterstainedwith Harris's haematoxylin, dehydrated,cleared, and mounted in DPX (BDH Ltd).
Sections were assessed by light microscopy,with regard to overall cellularity, distributionof granulocytic, erythroid, and megakaryo-cytic lineages, extent of maturation withineach lineage and presence of dysplastic cyto-logical features. For specimens taken beforetreatment, immunophenotypes of leukaemicinfiltrates were determined by counting thenumber of blast cells staining positively witheach antibody within a sample of 300 cells inrandomly selected fields.
ResultsIMMUNOPHENOTYPES OF PRETREATMENTBIOPSY SPECIMENSTable 2 summarises the immunophenotypesof the leukaemic infiltrates. CD45 and CD68,where positive, were expressed by all blastcells. Neutrophil elastase, CD 15, and cal-granulin were expressed by only a proportionof blasts cells in positive cases and theirexpression correlated loosely with FAB AMLsubtype. In particular, the presence of cal-granulin was associated with monocytic dif-ferentiation (AML-M4 and M5). Unexpectedreactivity with the antibody Ret40f (anti-flSGP) was found in some cases of otherwisetypical AML-M2 and one case of ALL.These cases were also stained using anothermonoclonal antibody, BRIC 4 (a gift from DrD Anstee, IBGRL, Southmead Hospital,Bristol), reactive with a different epitope ofSGP. Blast cells did not react with BRIC 4.The three cases in which AML developed
from myelodysplasia (all FAB subtype refrac-tory anaemia with excess blasts (RAEB) atinitial presentation) had AML phenotypes ofAML-M2, M4, and M6, respectively, by thetime frank leukaemia was present. Thepatient who developed AML-M4 had a tran-sient phase of RAEBt in which more than50% of blasts were erythroid: he received asingle course of DAT 3 + 10 at that timeand immediately afterwards had marrow indi-cators ofAML-M4.
CASES OF MINIMAL RESIDUAL DISEASEIn two cases (one AML-M2 and one AML-M4), residual leukaemic blast cells wereclearly demonstrated by immunohistochemi-cal staining of specimens (both at three weeksafter treatment). In one case (AML-M4)these had not been seen with routine stainsand in the other (AML-M2) they had beeninterpreted as equivocal, considered possiblyto be regenerating normal elements.
In the case of AML-M4 expression of cal-granulin and CD15 by large mononuclearcells represented a phenotype which charac-terised the patient's leukaemic blasts beforetreatment and which is found only in rare
916
on August 5, 2021 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.46.10.915 on 1 October 1993. D
Figures in parentheses denote weak positive staining reaction. = no positively staining blastcells; + = 95% or more of blast cells stained positively. Percentage values (to nearest 5%) aregiven for cases in which fewer than 95% of blast cells stained positively. AUL = acute undiffer-entiated leukaemia.Cases categorised according to the French-American-British (FAB) Group classification.
mononuclear cells in normal marrow biopsyspecimens.'4 In both cases, however, abnor-mal distribution of primitive cells and lack ofassociated maturing forms also contributed tothe diagnosis of residual leukaemia.
ORDER AND TIMING OF REGROWTHIn five cases marrow biopsy specimensshowed regenerative hyperplasia affecting dif-ferent combinations of erythrocytes, granulo-cytes, and megakaryocytes by the time of thefirst biopsy specimen taken after treatment(three to eight weeks after treatment). Thiswas unrelated to age (the patients were aged2, 10, 24, 25, and 44 years), diagnosis, ortherapeutic regimen. Not enough subsequentspecimens were available for adequate assess-ment of the duration of this regenerativeovershoot, but normocellular marrow wasfound in subsequent biopsy specimens fromthree affected individuals two, nine, and 16weeks later, respectively.
X, w. *; X :C,3§
wdi ,_ .*t: r._ .
wc _,":X.}.s s _., e * ^1 s t
e 4: v.. .:s 8 .t w.* k,X ' ;d
Figure 1 Erythroid cluster in which all cells seem to be at a similar stage of maturation("synchronous cluster"). In this example most of the cells are mid-normoblasts.Immunostained for /3-SGP using the antibody Ret4Of
CYTOLOGICAL DYSPLASIA AND ABNORMALITESOF SPATIAL ORGANISATIONCytological abnormalities were prominent inerythroid cells in 14 of 20 patients after treat-ment. One pattern of abnormality, seen innine patients, was characterised by lack of thenormal heterogeneity of red cell maturationstages within individual erythroid clusters, sothat each cluster seemed to be composed ofsynchronous cells (fig 1). Such clustersappeared denser and more cohesive than nor-mal, with sharply defined margins. Clustersof early cells coexisted in the same biopsyspecimen with others composed of late cellsbut mixed clusters were absent. In five othercases erythropoiesis appeared megaloblasticthroughout the entire biopsy specimen (fig 2).
Another abnormality of erythropoieticclusters concerned their spatial distribution.In normal marrow erythropoietic foci arenot found in the immediate vicinity oftrabeculae.15 In marrow biopsy specimenstaken after treatment from 12 of our 20patients, paratrabecular clusters of erythroidcells were prominent. These cases showed asharp reduction in granulopoiesis at a timewhen erythropoiesis had at least partiallyrecovered. In patients for whom later biopsyspecimens were available for evaluation, thedistribution of erythropoiesis returned to nor-mal as granulopoiesis recovered.
Dysplastic cytological features within gran-ulocytes were not seen, but a trephine biopsyspecimen is not a sensitive means of evaluat-ing such changes, especially if acid decalcifi-cation (with consequent loss of granularity)has been performed, as in our material. Themost usual abnormality detected in this studywas an overall reduction in the quantity ofgranulopoiesis, combined in some cases witha relative increase in numbers of immatureforms and paucity of mature granulocytes.No striking spatial abnormalities of granu-lopoiesis were seen; as regrowth proceeded,early and late granulocytes repopulated theirnormal sites.
Megakaryocyte dysplasia was prominent in14 of 20 cases, with abnormal clustering of'cells and pronounced abnormalities ofnuclear morphology in individual cells (fig 3).Irregularly segmented nuclei, mononuclearvariants, micromegakaryocytes and cells withbizarre angulated nuclei were found. Inpatients for whom biopsy specimens takenover a long period of follow up were available,these changes persisted for many weeks oreven months.
Monocytes are difficult to recognise in rou-tinely stained biopsy specimens of normalbone marrow, but can be shown byimmunostaining for calgranulin. They wereabsent, or present in only very low numbers,in specimens taken after treatment. LargeCD68 + macrophages containing abundantphagocytosed debris were common and per-sisted for many weeks in some patients. Theirpresence was independent of the speed ofregeneration of normal haemopoiesis or thepresence of dysplasia.
Mast cells were shown using the antibody
917
on August 5, 2021 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.46.10.915 on 1 October 1993. D
Figure 2 Megaloblast-like erythropoiesis. Irregular clusters of erythroid cells wmegaloblastic cytological appearances, immunostainedfor a-SGP using the antBRIC 101.
AA1 which recognises tryptase.12 Ifour specimens they were entirelywere present only in small number10 cells per intertrabecular space).the four exceptions were biopsy staken before treatment from paticALL. Biopsy specimnens taken after ifrom these patients did not contamast cell numbers. The other specintaining numerous mast cells (20 orintertrabecular space) were biopsy staken after treatment from two patiAML in whom the specimnens tak4treatment had contained only occasicells. None of these four patients shclonged hypoplasia after treatment. 1no correlation between the numbeicells present and the occurrence of 4changes in regenerating haemopoieti
In one patient pronounced stromand oedema were present throug
.......iS' ~ ~ GA
100:-.., A_m c
.q .. . _ .... ., S~~~~~~~~~~~~~~~~~~~~~~~~~~0
*
Figure 3 Megakaryocyte dysplasia highlighted by immunostainingfor CD61 using theantibody Y2/51. The megakaryocytes vary widely in size and shape, with severalmicromegakaryocytes evident in this field.
biopsy specimen four weeks after inductiontreatment, but good regrowth was presenttwo weeks later, with near normal peripheralblood cell counts.
CORRELATION OF BIOPSY FINDINGS WITHCLINICAL VARIABLES, PERIPHERAL BLOOD, ANDASPIRATE APPEARANCESThe level of erythropoiesis and granulopoiesisseen m trephine biopsy specimens correlatedwell with aspirate and peripheral blood find-
;:i8 ings. Megakaryocyte numbers, however, didnot show a consistent relation to bloodplatelet counts, which returned to normalwithin three to four weeks, regardless of thenumber of megakaryocytes in the biopsyspecimens.
Erythrocyte dysplasia in the biopsy speci-mens was accompanied by dyserythropoieticchanges in aspirate smears and by anaemia.
Pith The megaloblast-like changes seen in theribody specimens did not correlate with megalo-
blastic aspirate appearances or peripheralblood macrocytosis.
n all but Only two patients in our series hadabsent or received methotrexate; one of these hads (one to normoblastic erythropoiesis. Megaloblast-like. Two of change is therefore presumed to be a non-,pecimens specific toxic effect of various chemothera-ents with peutic agents.treatment There were no obvious differences betweeniin raised regrowth after chemotherapy for ALL rathernens con- than AML.more per Dysplastic changes seen after treatment;pecimens were not dependent on the presence ofients with known myelodysplasia preceding develop-en before ment of leukaemia.onal mast Regrowth seemed to be generally quicker)wed pro- in children and young adults than in older[here was adults, but analysis of age in relation to timer of mast of first normocellular biopsy specimen afterdysplastic chemotherapy did not reveal any significantic cells. difference between age groups.al fibrosis Suspected infections were treated promptlybhout the with intravenous antibiotics in all patients
and there was generally no detrimental effecton marrow regrowth. One patient had a seriesof infections which were difficult to control.His haemopoietic regrowth was slow (firstnormocellular biopsy specimen at 44 weeksafter induction therapy), but he had an unusu-
+ ally complicated clinical course. His ALL wasE;.>_ refractory to treatment. During treatment he
required surgery for debridement and graftingof a necrotising infected skin site.
Leukaemic relapse after treatmentoccurred in three patients and was unrelatedto good or poor initial haemopoietic regrowthin this study.Two patients received marrow autografts
after achieving remission. Biopsy specimenstaken shortly afterwards showed a variety ofhypoplastic and dysplastic changes, presum-ably reflecting toxicity of the conditioningregimen, which were indistinguishable fromthe effects seen in biopsy specimens taken
.::..:?..after primary treatment.
CASES OF DELAYED MARROW REGROWTHAmong our series of patients, four individuals
918
s
3.0.
on August 5, 2021 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.46.10.915 on 1 October 1993. D
Haemopoietic regrowth after chemotherapy for acute leukaemia
had exceptionally prolonged periods (morethan three months) of poor haemopoieticfunction as judged by blood, marrow aspirate,and trephine biopsy findings. One was thepatient with recurrent sepsis. The secondpatient had a diagnosis of AUL, subsequentlyrevised to ALL, and her disease respondedpoorly to both AML and ALL therapeuticregimens; her first normocellular biopsy spec-imen was obtained 27 weeks after startingtreatment. The third patient was a 54 yearold man with AML-M4, for whom no readilyidentifiable cause of poor regrowth wasfound. His disease subsequently relapsed,after which a marrow biopsy specimenshowed good regrowth with normal cellularityby six weeks. He sustained a further relapse,however, after autologous marrow transplan-tation 17 weeks later. The fourth patient wasa 49 year old man who developed hypo-cellular AML-M2 on a background ofmyelodysplasia. Despite autologous marrowtransplantation he had persistently hypoplas-tic marrow with prominent megakaryocyteand erythroid dysplasia up to 54 weeks fromstarting treatment.
DiscussionHaemopoietic regeneration after treatmentfor acute leukaemia is usually monitored byrepeated peripheral blood and bone marrowaspirate analyses. Our results show that thereis generally good correlation between theappearances in bone marrow biopsy speci-mens and those in concurrent blood or aspi-rated marrow samples over extended periodsof time after induction treatment. This find-ing confirms the validity of blood and aspiratesamples for monitoring bone marrow func-tion during and after treatment for acuteleukaemia.A study of biopsy samples, however, offers
the opportunity to evaluate aspects of the spa-tial organisation of haemopoietic cells, whichmay be of importance in determining the ade-quacy of reconstitution of normal haemopoi-etic function. Complex interactions betweenstromal cells, extracellular matrix molecules,growth factors and haemopoietic progenitorcells are required for normal production ofblood cells.'6 Abnormal spatial organisationof the bone marrow may reflect, or be a causeof, failure of this process. For example, spa-tial as well as cytological disturbances ofhaemopoietic cells are found in the primarymyelodysplastic syndromes and are frequentlyassociated with ineffective haemopoiesis. 17
The use of immunohistochemistry in con-junction with spatial and cytological analysisof marrow biopsy specimens permits cleardemonstration of the lineages and stages ofmaturation of the cells present.' '1
Because of the retrospective nature of ourstudy, the number and timing of biopsy sam-ples from different patients were variable sothat direct comparisons could not be madebetween all individuals. In keeping with pre-vious studies, however,' 2 our results showthat profound disturbances of haemopoiesisoccur after treatment for AML and ALL. The
availability of serial biopsy specimens fromour patients allowed us to demonstrate widevariation in the speed and sequence ofhaemopoietic regeneration. Relatively earlybiopsy specimens, taken between three andeight weeks from the start of treatment,showed the existence of a transient phase ofhyperplasia in a proportion of patients, thecause of which remains unexplained.Persistent hypoplasia, on the other hand,coexisted in all but one case with a compli-cated clinical course, albeit for different rea-sons in each patient.
Dysplastic changes took the form of cyto-logical and spatial abnormalities, predomi-nantly affecting erythroid cells andmegakaryocytes. Abnormal paratrabeculardistribution of erythrocytes was not necess-arily associated with failure of erythroid mat-uration, indicating that this area of bonemarrow stroma, normally occupied by earlygranulocytes, permits normal erythropoiesis.This suggests that marrow stromal compart-mentalisation may be less rigid than cell cul-ture derived models of specific haemopoieticmicroenvironments would imply.18 In ourbiopsy specimens repopulation of paratrabe-cular sites by granulocytes was accompaniedby return of erythropoiesis to its more usualsites in the central areas of marrow spaces,possibly due to simple physical displacement.The appearance of synchronous erythroid
colonies in specimens taken after treatmentmay reflect the development of new coloniesin a marrow which has been suddenly andseverely depleted of the normal mixture ofcells at different maturation stages. It is a fea-ture seen in some cases of primary myelodys-plasia"7 and could be the result of subtleabnormalities of interactions between stromaand developing erythroid cells. The mega-loblast-like changes seen in some biopsy spec-imens also resembled those found in somecases of primary myelodysplasia and did notcorrelate with peripheral macrocytosis. Inonly one patient were megaloblast-likechanges temporally related to treatment witha folate antagonist.
Megakaryocyte atypia was often severe andprolonged. There was no correlation betweenthe presence of cytologically atypicalmegakaryocytes and abnormal peripheralblood platelet counts. Micromegakaryocytesand hypersegmented megakaryocytes, indis-tinguishable from those found in primarymyelodysplastic"7 and myeloproliferativestates,'9 were seen. These abnormal appear-ances after chemotherapy may have been dueto nuclear damage sustained by long lived,non-dividing cells in which cytoplasmic func-tions, including platelet production, may berelatively normal.
Granulocyte dysplastic changes were notsevere, although prolonged granulocytehypoplasia occurred in some patients. Whengranulopoietic cells regrew, they did so attheir normal sites. This contrasts with pri-mary myelodysplasias, in which abnormallocalisation of granulopoietic cells is often aprominent feature. 7
919
on August 5, 2021 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.46.10.915 on 1 October 1993. D
There have been few previous studies oftrephine biopsy specimens taken duringmarrow regrowth after chemotherapy,' 2 20 andnone has used immunohistochemical stainingto confirm the lineages of regrowing cells.Two studies have reported biopsy findings inthe early period after the start of treatmentfor leukaemia.' 20 Essentially no regrowth waspresent in biopsy specimens taken at day 6and the prognostic value of the level of reduc-tion of leukaemic cells at this stage remainsuncertain.20 Up to four weeks after treatmentWittels' found little regrowth. Between fourand nine weeks, erythroid and megakaryocytenumbers typically recovered more quicklythan those of granulocytes.
Islam2 has described biopsy specimenstaken from 20 patients one to six weeks afterchemotherapy for AML had been completed.He found patchy regeneration at trabecularmargins and clustered centrally within inter-trabecular spaces, with gradual extensionfrom both sites to confluence. The lineages ofthe regenerating cells, however, were notdescribed, nor was the presence or absence ofdysplastic cytological features recorded.Two detailed studies, one retrospective
and one prospective, of trephine biopsy speci-mens taken after bone marrow transplanta-tion have been reported by van den Berg andcoleagues.2122 They found many features whichresemble those seen in the post-treatmentmarrow biopsy specimens taken after treat-ment and described here. In particular, theyobserved synchrony and megaloblasticchanges in erythroid cells and clustering ofmegakaryocytes. They found that granulo-poiesis generally recovered later than erythro-poiesis, but appearances were stable and nearnormal one month from transplantation.The variation in order or regrowth ofhaemopoietic cell lineages and the noticeablepersistence of some dysplastic features in ourstudy may reflect the different compositionand greater intensity of chemotherapeuticregimens used for initial treatment ofAMIJALL in comparison with pre-transplantconditioning treatment.
It has been suggested that high mast cellcounts may impede marrow regrowth follow-ing transplantation in cases of aplasticanaemia23 although another study found noevidence for this. We have already shownthat immunostaining for mast cell tryptase intrephine biopsy specimens, using the mono-clonal antibody AA1, reveals consistentlyhigher numbers of mast cells than doesmetachromatic staining.25 We were interestedto see whether mast cell numbers were associ-ated with poor regrowth in the context ofacute leukaemias. High mast cell numberswere found in only two specimens taken aftertreatment, and showed no association withpoor or dysplastic regeneration.
Stromal fibrosis has also been reported tohave an adverse effect on haemopoietic recov-ery after marrow transplantation,26 but therewas no delay in regrowth in the one patient ofour series in whom fibrosis was present afterinduction chemotherapy. Nor were persistent
abnormalities of spatial organisation or cyto-logical atypia found in this case, suggestingthat such fibrosis may be focal or transientand is not necessarily of clinical importance.Some of the cases reported by Wittels' alsodeveloped marrow fibrosis without noticeablyimpaired regrowth.Our study has also provided data of rele-
vance to the assessment of minimal residualdisease in acute leukaemia. In biopsy speci-mens taken as early as two or three weeksafter initial chemotherapy, granulopoiesis waseither completely absent or was representedby cells of all maturation stages. The quantityof granulopoiesis varied widely, but the onlybiopsy specimens showing focal clusters ofprimitive granulocytic/monocytic cells unac-companied by later forms came from patientswho had persisting leukaemic blasts in theirperipheral blood or marrow aspirate samples.Although these findings are circumstantial,they suggest that clusters of primitive lookingcells in biopsy specimens taken shortly aftertreatment should be interpreted with cautionand not dismissed as normal regeneration.
Immunohistochemical characterisation ofmyeloid leukaemic blast cells in trephinebiopsy specimens did not provide a sensitivemeasure of normal v leukaemic regenerationin this study, reflecting the lack of detectableantigenic differences between neoplastic cellsand early myeloid precursors. The rarity ofmonocyte precursors expressing calgranulinor CD 15 in normal marrow,'4 however,makes these antigens good potential targetsfor detection of small foci of AML-M4 orM5. A new antibody, PG-Mi, reactive withan epitope of CD68 different from that recog-nised by KP1, has been reported recently todiscriminate AML-M4 and M5 from otherFAB types and to show no reactivity withearly granulocytes.27 This may be of greatvalue in future for the assessment of residualor relapsed AML-M4/M5. Detailed analysisof lymphocyte associated antigens, not a pri-mary objective of this particular study, couldsimilarly provide helpful information for thediagnosis of residual ALL.
1 Wittels B. Bone marrow biopsy changes followingchemotherapy for acute leukemia. Am Y Surg Pathol1980;4:135-42.
2 Islam A. Pattern of bone marrow regeneration followingchemotherapy for acute myeloid leukemia. J Med1987;18: 108-22.
3 Wilkins BS, Jones DB. The use of immunohistochemicaltechniques with fixed, decalcified, wax-embedded bonemarrow trephine biopsies. Haematol Rev 1989;3:149-63.
4 Van der Valk P, Mullink H, Huijgens PC, Tadema TM,Vos W, Meijer CJL. Immunohistochemistry in bonemarrow diagnosis: value of a panel of monoclonal anti-bodies on routinely processed bone marrow biopsies.Am J Surg Pathol 1989;13:97-106.
5 Caldweli CW, Patterson WP, Toalson BD, Yesus YW.Surface and cytoplasmic expression of CD45 isoformsin normal and malignant myeloid cell differentiation.AmY Clin Pathol 1991;95:180-7.
6 Pulford KAF, Rigney EM, Micklem KJ, et al. KPI: a newmonoclonal antibody that detects a monocyte/macrophage associated antigen in routinely processedtissue sections. J Clin Pathol 1989;42:414-21.
7 Pulford KAF, Erber WN, Crick JA. Use of a monoclonalantibody against human neutrophil elastase in normaland leukaemic myeloid cells. J Clin Pathol 1988;41:853-60.
8 Albrechtsen M, Kerr MA. Characterisation of humanneutrophil glycoproteins expressing the CD15 differen-tiation antigen (3-fucosyl-N-acetyllactosamine). Br JHaematol 1989;72:312-20.
920
on August 5, 2021 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.46.10.915 on 1 October 1993. D
11 Gatter KC, Cordell JL, Turley H, et al. The inumuno-histological detection of platelets, megakaryocytesand thrombi in routinely processed specimens.Histopathology 1988;13:257-67.
12 Walls AF, Jones DB, Williams JH, Church MK, HolgateST. Immunohistochemical identification of mast cellsin formaldehyde fixed tissue using monoclonalantibodies specific for tryptase. J Pathol 1990;162:119-26.
13 Hsu S-M, Raine L, Fanger H. Use of avidin-biotin-perox-idase complex (ABC) in immunoperoxidase techniques:a comparison between ABC and unlabelled antibody(PAP) procedures. J Histochem Cytochem 1981;29:577-80.
14 Wilkins BS, Jones DB. Cell-stroma interactions in mono-cytopoiesis. FEMS Immunol Microbiol 1992;105:347-54.
15 Wilkins BS. The histology of normal haemopoiesis: bonemarrow histology I. J Clin Pathol 1992;45:645-9.
20 Roberts MM, Juttner CA, To L Bik, Kimber RJ. Bonemarrow biopsy during induction chemotherapy foracute myeloid leukaemia identifies only 50% of patientswith resistant disease. Leuk Res 1988;12:817-21.
21 Van den Berg H, Kluin PM, Zwaan FE, Vossen JM.Histopathology of bone marrow reconstitution afterallogeneic bone marrow transplantation. Histopathology1989;15:363-73.
22 Van den Berg H, Kluin P M, Vossen J M. Early reconsti-tution of haematopoiesis after allogeneic bone marrowtransplantation: a prospective histopathological study ofbone marrow biopsy specimens. J Clin Pathol 1990;43:365-9.
23 Naeim F, Smith GS, Gale RP, and the UCLA BoneMarrow Transplant Team. Morphologic aspects ofbone marrow transplantation in patients with aplasticanemia. Hum Pathol 1978;9:295-308.
24 Sale GE, Marmont P. Marrow mast cell counts do notpredict bone marrow graft rejection. Hum Pathol1981;12:605-8.
25 Nock IS, Wilkins BS, Jones DB. Bone marrow mast cellsrevisited. J Pathol 1991;164:352A.
26 Rajantie J, Sale GE, Deeg HJ, et al. Adverse effect ofsevere marrow fibrosis on hematologic recovery afterchemoradiotherapy and allogeneic bone marrow trans-plantation. Blood 1986;67:1693-7.
27 Falini B, Flenghi L, Pileri S, et al. PG-MI: a new mono-clonal antibody directed against a fixative-resistant epi-tope on the macrophage-restricted form of the CD68molecule. AmY Pathol 1993;142:1359-62.
921
on August 5, 2021 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.46.10.915 on 1 October 1993. D
