Materials and Methods

Peripheral Blood Morphologic Changes after High-Dose Antineoplastic Chemotherapy and Recombinant Human Granulocyte Colony-Stimulating Factor Administration DANIEL P. KERRIGAN, M.D., ANNIE CASTILLO, M.D., KATHY FOUCAR, M.D., KELLY TOWNSEND, MT(ASCP), AND JAMES NEIDHART, M.D. The peripheral blood morphologic findings in 17 patients with cancer who had received high-dose cytotoxic chemotherapy followed by recombinant human-granulocyte colony-stimulating factor (rh-gcsf) were reviewed and compared with a control group of patients who received only high-dose chemotherapy. Both groups showed dysmyelopoiesis (abnormal granulation and nuclear lobulation) in the granulocytic series during the period of bone marrow recovery that followed the cytotoxic chemotherapy. Most of these morphologic abnormalities were more prominent in the rh-gcsf-treated group. Monocytic cells in both groups showed prominent vacuolation and immature nuclei. The percentages and absolute numbers of large granular lymphocytes were increased in the rh-gcsf group compared with the control group. No quantitative or qualitative abnormalities of eosinophilic or basophilic granulocytes were detected in either group. Both groups showed nonspecific red blood cell abnormalities, and large platelets were present in half of the control group smears. This report provides the first detailed peripheral blood morphologic description in patients treated with rh-gcsf and high-dose chemotherapy. (Key words: GCSF cytotoxic therapy; WBC morphology) Am J Clin Pathol 1989;92:280-285 A DOSE-LIMITING FACTOR in the treatment of human malignancies is neutropenia because of its association with increased risk of infection. 6 The neutropenia results from the direct toxic effects of antineoplastic therapy on granulocyte precursors in the bone marrow. 2 Therapy aimed at reducing this side effect has tremendous potential for decreasing morbidity and mortality in patients with cancer. One approach that has been proposed to counteract chemotherapy-induced myelosuppression is administration of hematopoietic growth factors. 3 A number of hematopoietic growth factors have recently been identified, sequenced, biosynthesized, and used in a variety of in vitro and in vivo studies. 3 One of these factors, recombinant human-granulocyte colony-stimulating factor (rh- GCSF), has been shown to primarily stimulate the production and enhance the function of granulocytes."" 13 Recent clinical trials using rh-gcsf have shown a significant decrease in neutropenia-associated complications Received October 27, 1988; received revised manuscript and accepted for publication January 30, 1989. Address reprint requests to Dr. Kerrigan: Department of Pathology, University of New Mexico, Albuquerque, New Mexico 87131. Departments of Pathology and Hematology/Oncology, University of New Mexico, Albuquerque, New Mexico when it has been given to cancer patients also receiving antineoplastic chemotherapy. 5,8 ' 9 This simultaneous and/ or sequential exposure of the bone marrow to a potent hematopoietic growth factor(s) and a potent toxin(s) was previously seen only in chemotherapy-treated patients with tumors producing endogenous hematopoietic growth factors. 10 The peripheral blood morphologic characteristics in this unique setting have been described only briefly. 8 In this article, we present a detailed description of sequential peripheral blood hemogram data and morphologic characteristics from 17 patients who received both high-dose chemotherapy and rh-gcsf. The relationship of these findings to the sequential bone marrow exposure to potent toxins and growth factors is discussed. Materials and Methods Sequential peripheral blood samples from 17 patients (10 male/7 female), ranging in age from 26 to 64 years, who were treated with rh-gcsf and high-dose cytotoxic chemotherapy for a variety of malignancies (10 carcinomas, 5 lymphomas, 1 melanoma, and 1 sarcoma), constitute this study. Fifteen of the 17 had received prior radiation and/or chemotherapy. The cytotoxic chemotherapy, consisting of cyclophosphamide (1.5-3.0 g/m 2 ), etoposide (350-600 mg/m 2 ), and cisplatinum (35-60 mg/ m 2 ), was administered during the time period shown in Figure 1. Decadron was given as an antiemetic during the first three days of treatment. Recombinant human- GCSF was given in doses of 20, 40, or 60 /tg/(kg day). The clinical and therapeutic details of a larger group of patients that included these 17 are reported elsewhere. 9 Eleven patients in this series received more than one cycle containing rh-gcsf (7 received two cycles and 2 received three cycles). Peripheral smears from nine high-dose chemotherapy cycles in seven patients who did not receive rh-gcsf were also reviewed. These findings were com- 280

Vol. 92 No. 3 GCSF AND CYTOTOXIC THERAPY EFFECTS ON WBCs 1000 E 281 FIG. 1. Individual patient total WBC count during an rh-gcsf-containing treatment cycle. Patientreceived60 /ig/kg/day of rh-gcsf. 100 WBC xiovi 10 '_ Cytotoxlo Chemotherapy rh-gcsf 1 0.1 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 Day of Cycle pared with the rh-gcsf-containing cycles to determine which morphologicfindingswere attributable to addition of rh-gcsf to the therapy protocol. Coulter S+IV (Coulter Electronics, Hialeah, FL) results, including white blood cell (WBC) count, platelet count, hematocrit, hemoglobin, mean corpuscular volume (MCV), MCHC, MCH, and red blood cell distribution width (RDW) were obtained daily on the patients' peripheral blood drawn in the morning. One hundred-cell differential counts were performed daily by technologists on Wright's-stained peripheral blood smears that generally were made within two hours of drawing the patient's blood. Peripheral smears from the day with the peak WBC count during the rh-gcsf administration were selected for more detailed morphologic review, and 200-cell differential counts were performed on these smears by one of the authors (D.P.K.). For patients who did not receive rh-gcsf, the peripheral smear from the day with the highest WBC count between days 7 and 28 was reviewed. In addition to differential counts, the presence and/or degree of toxic granulation, Dohle bodies, monocyte vaciiolation, dyspoietic changes, 1 large granular lymphocytes (LGLs), and red blood cell (RBC) and platelet abnormalities were noted on each blood smear. Results The most common pattern of WBC, granulocyte, lymphocyte, and monocyte fluctuation for patients receiving high-dose chemotherapy followed by rh-gcsf rescue is 1000 100 WBC xio*vi Granulocytes lymphocytes Monocytes -B- Immature Myeloid FIG. 2. Absolute cell counts for the cell cycle shown in Figure 1. 0.01 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 Day of Cycle

282 KERRIGAN ET AL. A.J.C.P.-September 1989 30 25 monocytes promyelocytes/blasts \m\ meta/myelocytes EH3 lymphocytes 20 - % of total WBQ 15 10 ^ I 1 I I m I I i I i 1 p I M PK DAY PK DAY PK DAY Day Of PK DAY PK DAY PK DAY -e -4-2 peak WBC *z *4 +e FIG. 3. WBC percentages before and after the peak WBC count observed during rh-gcsf therapy. Median values for the 25 cycles with leukocytosis. (Band and segemented neutrophils account for the remaining cell percentages.) shown in Figures 2 and 3. Three discrete WBC peaks and one period of neutropenia were seen. ThefirstWBC peak occurred during administration of Decadron and resulted from an increase in segmented neutrophils. A second smaller transient increase in WBC count was observed on day 7 in 70% of cycles. It occurred within 24 hours of the initial dose of rh-gcsf, lasted less than 48 hours, and went as high as 7 X 10 9 /L. This peak was composed of band and segmented neutrophils and was more common and pronounced in patients who had received rh-gcsf in prior cycles. This brief WBC elevation was followed by a period of profound neutropenia (granulocytes usually <0.1 X 10 9 /L) that lasted for 5-19 days (median, 9 days). In most patients, a brisk neutrophilia and monocytosis followed this period of neutropenia, developing between 18 and 28 days (median, 25 days) after the initiation of chemotherapy. The peak WBC during this period exceeded normal limits (10 X 10 9 /L) in 25 of 35 (78%) treatment cycles in which rh-gcsf therapy was administered. In contrast, the WBC count in patients treated with highdose chemotherapy alone never exceeded 10 X 10 9 /L, and no prominent WBC count peaks were seen. Patients who received the highest doses of rh-gcsf (60 /ug/kg/ day) had the most striking WBC responses. Peripheral Blood Morphology Band and Segmented Neutrophils (PMNs). Differential counts from the peak WBC count day during rh-gcsf treatment showed that most (>80%) of the WBCs were band or segmented neutrophils (Fig. 3). The median peak neutrophil count was 30 X 10 9 /L for the 32 cycles containing rh-gcsf. In every cycle in which rh-gcsf was administered,-including cycles in which the maximum WBC count was less than 10 X 10 9 /L, at least 25% of the bands and segmented neutrophils contained toxic granules and/or Dohle bodies. These changes were seen with similiar frequency in patients receiving all three dose levels of rh-gcsf. The number of toxic granules and Dohle bodies per neutrophil varied widely, but all smears contained easily identifiable cells with large numbers of both. In addition, cells in which most of the neutrophil cytoplasm remained moderately basophilic without distinct Dohle bodies were seen in smears with the most striking Dohle bodies. In contrast, hypogranular neutrophils were also present in about half the cases, generally accounting for less than 25% of the cells, although these cells predominated in one patient's smear. In some cases a combination of hypogranularity and Dohle bodies was present in the same cell (Fig. 4A). Neutrophils with increased diameter were also identified easily. Nuclear lobulation of granulocytes was abnormal in all cases with both hypersegmentation and hyposegmentation (Figs. 4B-D). In most smears at least 25% of the nuclei exhibited easily recognizable nuclear abnormalities. Nuclei with five or more lobes, psuedo-pelger-huet nuclei, and corkscrewshaped nuclei were the most common abnormalities, whereas circular nuclei were also seen (Figs. 4E and F). Slides from the bone marrow recovery period in the non-rh-gcsf cycles (days 17-34) showed many of the same findings as the rh-gcsf smears (Table 1), but in general the findings were seen in fewer cells and were much less prominent when present. Exceptions to this

Vol. 92 No. 3 GCSF AND CYTOTOXIC THERAPY EFFECTS ON WBCs 283 FlG. 4 (left). Peripheral blood granulocyte morphologic characteristics after high-dose cytotoxic chemotherapy and rh-gcsf. A. Hypogranular cytoplasm with Dohle bodies (arrows). B. Hyperlobulated nucleus. C. Hypolobulated nucleus. D. Circular nucleus. E. Convoluted "corkscrew" nucleus. F. Hypolobulated nucleus. Wright (X 1,000). FIG. 5 (right). Peripheral blood leukocyte morphologic characteristics after high-dose cytotoxic chemotherapy and rh-gcsf. A. Immature myeloid cell with blast-stage nuclear myelocytic-stage cytoplasmic maturation. B. Vacuolated monocyte. C. Mature and immature monocytoid cells. D. Lymphocyte with abnormal nuclear lobulation. E. Myeloid cell fragment. Wright (X 1,000). included hypogranular PMNs with Dohle bodies and hyperlobulated PMNs that were seen only in patients receiving rh-gcsf. The differential counts in these nonrh-gcsf cycles showed a lower percentage of granulocytes compared with the rh-gcsf-containing cycles (median percentages, 53% vs. 85% band plus segmented neutrophils, respectively). Immature Myeloid Cells (blasts through metamyelocytes). Immature myeloid cells averaged 2-10% of the WBC count during the rh-gcsf-induced peak WBC count period (Fig. 3). The highest absolute numbers of these immature cells occurred simultaneously with the peak WBC (Fig. 2). Myelocytes and metamyelocytes were more numerous than blasts and promyelocytes (Fig. 3). The percentages of immature myeloid cells appeared to be independent of the total WBC count. The highest absolute immature myeloid cell count was 35 X 10 9 /L. In cycles not containing rh-gcsf, the percentage, but not absolute number, of immature cells was slightly higher (median, 12.5%) during this period, with a much wider range (0.5-36%). Similar to the rh-gcsf-treated group, the meta/myelocytes outnumbered the promyelocytes/ blasts. The highest absolute immature myeloid cell count was 2.1 X 10 9 /L. Immature cells gradually disappeared from the blood as the WBC count returned to normal in both groups (Fig. 3). Dysynchrony between nuclear and cytoplasmic maturation was easily identified in most smears of the rh- GCSF-treated group. The predominant abnormality was a chromatin pattern and nuclear shape that was less differentiated than the cytoplasm (Fig. 5^4). In addition, very intense staining of the primary granules was seen in promyelocytes. Similar abnormalities were seen in the cycles not containing rh-gcsf, but they were less striking than in the rh-gcsf-containing cycles. Monocytes. The fluctuations in monocyte counts in patients receiving rh-gcsf closely resembled those of the granulocytes, and 22 of the 24 monocyte peaks occurred within 48 hours of the peak granulocyte count. Peak absolute monocyte counts ranged from 0.1 to 14 X 10 9 /L (median, 3.7 X 10 9 /L), and monocytes exceeded 1.0 X 10 9 /L in 24 of 35 rh-gcsf-containing cycles. The monocyte morphologic characteristics were remarkable for consistent prominent vacuolation in every case (Fig. 5B), even if an absolute monocytosis was not present. In addition, immature monocytes with moderately increased nuclear/cytoplasmic ratios, a fine chromatin pattern, and prominent nucleoli were seen in two-thirds of the cycles (Fig. 5C). The chemotherapy cycles lacking rh-gcsf showed identical morphologic findings in the monocytes. The absolute number of monocytes, however, was increased in only one of nine cycles. Lymphocytes. In rh-gcsf-treated patients, lymphocyte morphologic characteristics were generally unremarkable, although the proportion of large granular lym-

284 KERRIGAN ET AL. A.J.C.P. September 1989 Table 1. Peripheral Blood Morphologic Findings after High-Dose Chemotherapy With and Without rh-gcsf identified in patients receiving only high-dose chemotherapy. Peripheral Blood Morphologic Findings During Bone Marrow Recovery Toxic granulationf Dohle bodiesf Abnormal myeloid nuclear lobulation! Hypogranular PMNs present Percentage neutrophils >80% Vacuolated monocytes! Large granular lymphocytes (> 50% of lymphs) Immature monocytes (>1% of cells) Large platelets Myeloid cell fragments Immature forms of undetermined lineage (>1% of cells) Plasmacytoid lymphocytes (>1%) High-Dose Chemotherapy Cycles With rh-gcsf* 84% (both hyperlobulated and hypolobulated) 55% 66% 66% 5% Occasional 60% 12% * Percentage of cycles in which morphologic abnormality detected. t Findings were present in 25% or more of appropriate cells. Without rh-gcsf* 55% 22% 33% (hypolobulated only) Rare 22% 0 22% 77% 55% 66% 44% phocytes (LGLs) was greatly increased (>50% of lymphocytes) in two-thirds of these cases and the absolute number of large granular lymphocytes exceeded 1.0 X 10 9 / L in one-third of cases. In addition, lymphocytes with clover leaf and other nuclear abnormalities were seen on two smears. One of these patients was being treated for lymphoma and one for carcinoma (Fig. 5D). Two patients with carcinoma had small numbers (1-5%) of plasmacytic/immunoblastic lymphocytes with normal nuclear contours. Mildly elevated lymphocyte counts (3.6-5.0 X 10 9 /L) occurred within 24 hours of the peak WBC count in 24% of cycles with leukocytosis and in no cycle without leukocytosis. Review of smears from non-rh-gcsf-containing cycles showed an increased percentage of large granular lymphocytes in 22% of cases, but the absolute number of these cells never exceeded 1.0 X 10 9 /L. Abnormal lymphoid nuclear lobulation was seen in one cycle. Fortyfour percent of these cycles had 1% or more plasmacytoid/ immunoblastic lymphocytes, and in one case the cells contained Dutcher bodies. No lymphocyte elevations were Other Features Two additional abnormal features (cytoplasmic fragments and immature cells of undetermined lineage) were noted. The cytoplasmic fragments (ranging up to 15 nm in diameter) were seen in several smears from cases with WBC counts greater than 100 X 10 9 /L (Fig. 5E). The fragments had a faintly basophilic cytoplasm and varying numbers of strongly azurophilic granules. These fragments closely resembled the cytoplasm of immature granulocytic cells in the same smears. The immature cells of undetermined lineage were seen in half of the cases, always accounted for 5% or less of the cells, and tended to be more frequent in patients with higher WBC counts. They were similar in size, nuclear to cytoplasmic ratio, and chromatin pattern to promyelocytes but lacked primary granules. Some had partially indented nuclei, similar to an immature monocyte. The cytoplasm of these cells was faintly to moderately basophilic and occasionally contained vacuoles. These fragments and cells were seen in cycles with and without rh-gcsf but were more common in the rh-gcsf-containing cycles. Eosinophils and basophils always accounted for less than 1% of cells, and no morphologic abnormalities were noted. Except for one case that showed prominent basophilic stippling, red blood cell morphologic changes were limited to mild to moderate nonspecific anisopoikilocytosis, mild polychromasia, and occasional (<2 per 100 WBCs) nucleated red blood cells. Abnormal platelet morphologic characteristics were limited to hypogranular platelets in one rh-gcsf-treated patient and large platelets (>7 ^m diameter) in half of the control group patients. Discussion Granulocyte colony-stimulating factor and other hematopoietic growth factors have recently become available in purified form through the efforts ofscientists in a variety of fields. 3 DNA cloning of the human GCSF gene, accomplished in 1986, has allowed production of sufficient GCSF for a variety of in vitro and, more recently, in vivo studies of its effect on hematopoiesis. 5 The preliminary in vitro and in vivo data have shown that GCSF is a potent stimulator of granulocyte function and production. It has also been shown to reduce cytotoxic chemotherapy-associated complications in humans. 5,813 This is the first study to provide a detailed description of the peripheral blood morphologic characteristics in patients receiving intensive therapy with antineoplastic chemotherapy followed by rh-gcsf. The morphologic

Vol. 92 No. 3 GCSF AND CYTOTOXIC THERAPY EFFECTS ON WBCs 285 changes seen in these cases reflect the combined effects of the antineoplastic drugs and a potent hematopoietic growth factor (rh-gcsf). We suggest the morphologic changes attributable to rh-gcsf administration are the hyperlobulated PMNs and the increased toxic granulation and Dohle bodies in neutrophils. 8 Myelodysplastic changes, including abnormal nuclear lobulation, hypogranular cytoplasm, and nuclear-cytoplasmic maturation dysynchrony, are well described after antineoplastic therapy, 4 and the antineoplastic therapy probably accounts for most of these findings in our cases. It is the combination of these findings (along with the significant leukocytosis) in both a single patient and within a single cell that seems unique to this patient population. The more pronounced myelodysplastic changes in the rh-gcsftreated patients suggests that rh-gcsf stimulated myeloid cell proliferation and maturation is occurring at a point when the marrow has not recovered from the cytotoxic chemotherapy injury. The cause of the additional peripheral blood abnormalities, including monocyte vacuolation, WBC fragmentation, increased absolute numbers of large granular and/or plasmacytoid lymphocytes, and occasional abnormal lymphocyte nuclear lobulation, is uncertain. The immunophenotypic and functional characteristics of the LGLs may be of clinical interest because tumor-killing lymphocytes can morphologically be LGLs. It may also be interesting to examine molecular and/ or cytogenetic events associated with the striking morphologic changes seen in this setting because many of these abnormalities are those seen in myelodysplastic syndromes 4 and chemotherapy-related leukemias. 7 Acknowledgments. The authors thank L. Sousa, Ph.D., and M. Downing, Ph.D., of AMGEN Corporation of Thousand Oaks, California, for reviewing the manuscript and supplying the rh-gcsf used in this study. References 1. Bennett JM, Catovsky D, Daniel MT, et al. Proposals for the classification of the myelodysplastic syndromes. Br J Haematol 1982;51:189-199. 2. Brunning RD. The effects of leukemia and lymphoma chemotherapy on hematopoietic cells. Am J Med Technol 1973;39:165-174. 3. Clark SC, Kamen R. The human hematopoietic colony-stimulating factors. Science 1987;236:1229-1237. 4. Foucar K., McKenna R, Bloomfield C, et al. Therapy related leukemia, a panmyelosis. Cancer 1979;43:1285-1296. 5. Gabrilove JL, Jakubowski A, Scher H, et al. Effect of granulocyte colony stimulating factor on neutropenia and associated morbidity due to chemotherapy for transitional-cell carcinoma of the urothelium. N Engl J Med 1988;318:1414-1421. 6. Hoagland HC. Hematologic complications of cancer chemotherapy. Semin Oncol 1982;9:95-101. 7. Kantarjian HM, Keating MJ. Therapy-related leukemia and myelodysplastic syndrome. Semin Oncol 1987;14:435-443. 8. Morstyn G, Souza L, Keech J, et al. Effect of granulocyte colony stimulating factor on neutropenia induced cytotoxic chemotherapy. Lancet 1988;1:667-671. 9. Neidhart J, Kohler B, Castillo A, et al. Recombinant human granulocyte-colony stimulating factor (rhg-csf) shortens duration of granulocytopenia and thrombocytopenia following intensive chemotherapy dosing. (Submitted for publication). 10. Robinson WA. Granulocytosis in neoplasia. Ann NY Acad Sci 1974;230:212-217. 11. Souza LM, Boone TC, Gabrilove J, et al. Recombinant human granulocyte-colony stimulating factor: effects on normal and leukemic myeloid cells. Science 1986;232:61-65. 12. Welte K, Bonilla MA, Gabrilove JL, et al. Recombinant human granulocyte-colony stimulating factor: in vitro and in vivo effects on myelopoiesis. Blood Cells 1978;13:17-30. 13. Yuo A, Kitagawa S, Okabe T, et al. Recombinant human granulocyte-colony stimulating factor repairs the abnormalities of neutrophils in patients with myelodysplastic syndromes and chronic myelogenous leukemia. Blood 1987;70:404-411.