Splenic Myeloid Metaplasia, Histiocytosis, and Hypersplenism in the Dog (65 Cases)

Transcription

1 Vet Pathol 36: (1999) Splenic Myeloid Metaplasia, Histiocytosis, and Hypersplenism in the Dog (65 Cases) W. L. SPANGLER AND P. H. KASS Anatomic Pathology Section, IDEXX Veterinary Services Inc., West Sacramento, CA (WLS); and Department of Population Health and Reproduction, School of Veterinary Medicine, University of California, Davis, CA (PHK) Abstract. Splenectomy specimens from 65 dogs with severe, diffuse, sustained, and progressive splenomegaly were examined. The clinical signs, hematology, and serum chemistry values in for the dogs were not useful diagnostic features. Microscopic changes in the spleens were distinctive and consisted of 1) myeloid metaplasia, 2) histiocytosis, 3) erythrophagocytosis, and 4) thrombosis with segmental infarction. Ultrastructural features suggested proliferative changes in the splenic reticular cells and macrophages (reticular meshwork) that described a continuum from reactive changes associated with immunologic damage of erythrocytes to neoplastic proliferation of histiocytic components. Thirty percent of the dogs survived 12 months. Approximately one half (53%) of the dogs with complete postmortem evaluations showed multiorgan involvement with a tissue distribution and cell morphology consistent with histiocytic neoplasia. For the remaining dogs (47%), only splenic pathology was consistently present, and a specific cause of death was often not evident. Distinctive histologic changes in the splenic tissues including mitotic activity, erythrophagocytosis, giant cell formation, thrombosis/ infarction, and the proportion and distribution of histiocytic and hematopoietic cells were statistically evaluated for prognostic relevance. The presence of giant cells was the only reliable prognostic feature, and that was indicative of a fatal outcome. These descriptive changes of myeloid metaplasia in the canine spleen are compared with the human clinical and pathologic syndromes of 1) agnogenic myeloid metaplasia, 2) hemophagocytic syndromes, and 3) hypersplenism. These diseases in humans produce histopathologic changes in the spleen that are similar to those observed in the canine splenic tissue we examined in this study. Key words: Canine species; electron microscopy; fibrohistiocytic cell; hemophagocytic syndrome; hypersplenism; myeloid metaplasia; patient survival; spleen. Splenomegaly in the dog is most often encountered pathologically as asymmetric regional nodular distortion. 15,16 Uniform or more symmetrical splenomegaly may occur from diffuse cellular proliferation or infiltration of the spleen and may be seen in conjunction with a wide variety of clinical illnesses. 13 Pathologic changes in these infiltrated or reactive spleens are often not specifically diagnostic and may suggest a clinical diagnosis of histiocytosis, rickettsial or protozoan disease, lymphoma, chronic bacterial infection, or autoimmune hemolytic anemia. 13 Splenomegaly, in these cases, is often the physical and diagnostic centerpiece of the patient s clinical illness. A section or aspirate of the enlarged spleen is usually the diagnostic specimen submitted to define both the disease process and the appropriate therapy. This paper describes a distinctive form of splenic pathology characterized grossly by diffuse, uniform, persistent, and progressive splenomegaly. Clinically, this type of spleen is typically derived from a dog with vomiting, lethargy, fever, malaise, and cytopenia (usually anemia and/or thrombocytopenia) and may be associated with a variety of clinically distinct illnesses with a range of treatment regimens. Histologically, the spleens are distinctive and can be recognized by 1) diffuse hematopoiesis (myeloid metaplasia), 2) vascular thrombosis with segmental splenic infarction, and 3) reactive and proliferative cellular reticular meshwork containing excess macrophages. The macrophages represent a continuum from reactive histiocytosis with 4) excessive erythrophagocytosis in the bone marrow and spleen (hypersplenism) to 5) overtly malignant histiocytic neoplasms 4,11,20 22 with giant cells and multiorgan involvement. These pathologic changes in the canine spleen are compared with similar conditions described in human pathology: 1) agnogenic myeloid metaplasia, 19,22 2) hemophagocytic syndrome, 1,2 and 3) hypersplenism. 22 All of these conditions are associated with cytopenias. Both hypersplenism and hemophagocytic syndrome are considered important spectral components of histiocytosis syndromes involving ordinary phagocytic macrophages found in the splenic cords. 1 Individual portions of this spectrum of histiocytic disorders af- 583

2 584 Spangler and Kass Vet Pathol 36:6, 1999 fecting the canine spleen can be identified in the literature as single-case reports or reports of small groups described as distinct entities. 4,6,8,11,14 Myeloid metaplasia, as a pathologic diagnosis applied to the canine spleen, has been documented in previous work. 21 Unfortunately, correlation of this arcane splenic lesion with the associated clinical diseases, pathologic presentation, and prognosis has thus far eluded clinical and pathologic insight. Because the cases of canine splenomegaly described in this paper share common pathologic features as well as a poor clinical outcome, they appear to provide a useful diagnostic category with prognostic relevance. The prevalence of canine splenic lesions and their serious consequences in aging dogs, along with the spleen s surgical and diagnostic accessibility (via palpation and ultrasound), make the spleen a good candidate organ for regular periodic clinical evaluation in aging dogs as part of the health maintenance and preventive medicine environment of current veterinary practice. Materials and Methods The 65 specimens included in this report were selected from surgically and necropsy-derived splenic tissues submitted to the laboratory (CVD-IDEXX Veterinary Services, West Sacramento, CA 95605) during a 10-year period ( ) and represent 1.5% of approximately 4,000 spleens examined during that period. Inclusion of spleens in the study was dependent on 1) splenomegaly, 2) the histologic appearance of the tissue (inappropriate hematopoiesis/ myeloid metaplasia), and 3) the availability of postoperative survival information on the dog. A gross description of the spleen was available for 51% (33/65) of the dogs. Complete postmortem examinations of dogs that died subsequent to splenectomy were performed (by WLS) on 23% (15/65) of the dogs, and electron microscopy (initial fixation of splenic sections in gluteraldehyde) was available for 18% (12/65) of the dogs. Selected hematology and/or clinical chemistry data were available for 89% (58/65) of the dogs. The data were obtained from preand postoperative profiles available in the laboratory files or the veterinarian s clinical records and ranged from 150 days prior to splenectomy (dog No. 18) to 30 days postsplenectomy (dog No. 65). The median and mean times of data collection were 3 days and 1 day (prior to splenectomy), respectively. Patient survival information was obtained for all dogs from veterinarians who completed a questionnaire that included the date of death or euthanasia for all dogs along with the cause of death or reason for euthanasia, if known. Dogs were considered to have been euthanatized or to have died as a direct result of their splenic disease if 1) they died in the immediate postoperative period ( 15 days following splenectomy); 2) metastatic lesions were present at the necropsy exam; or 3) the referring veterinarian could, through the use of ancillary clinical procedures (i.e., radiographs, fine-needle aspirate, palpation, etc.), be reasonably certain that death or euthanasia was brought about by complications associated with the splenic disease. Follow-up data were obtained on all dogs for a minimum of 12 months postsplenectomy but ranged from 12 to 48 months (mean 24 months). Splenic tissue from all cases was processed by routine paraffin infiltration methods of histology and examined microscopically as 4 6- m hematoxylin and eosin (HE)- stained sections on glass slides. Specimens were prepared for transmission electron microscopy by mincing sections of fresh tissue in cold 2.5% gluteraldehyde. Five individual fragments were processed and embedded in 50% Bojax plastic/propylene oxide, sectioned at 1 m, and stained with toluidine blue. Areas with representative cell types were selected by light microscopy from three to five blocks, thinsectioned, mounted on copper grids, stained with lead citrate/uranyl acetate, and examined in a ZEISS EM 10A TEM (California Regional Primate Research Center, Davis, CA 95616). Distinctive histologic changes present in the tissues (e.g., erythrophagocytosis, giant cells, thrombosis/infarction, and the proportion and distribution of histiocytic and hematopoietic cells) within each spleen were tabulated and evaluated for diagnostic and prognostic relevance. Available clinical hematology, serum chemistry, and immunology data were similarly analyzed. The distribution of covariates was evaluated using Wilk s W statistic for normality and by construction of 95% frequency intervals using the 2.5th and 97.5th percentiles as end points. Data are also summarized by the mean, standard deviation, and median. The potential effects of the covariates on survival time were evaluated using a Cox proportionalhazards regression model. The Kaplan-Meier (product-limit) method of survival function estimation was used to estimate median survival time and 12-month survival probability. Results Patients with splenic myeloid metaplasia showed a slight gender bias: 57% (37/65) were female, and 43% (28/65) were male (1.3 : 1). Patient age (63/65) varied from 2 to 15 years (mean/median, 8 years). A breed was specified in each case: 65% (42/65) of all cases were accounted for by 15 Golden Retrievers, 6 Labrador Retrievers, 6 Shepherd-type dogs, 5 Rottweilers, 2 Chows, 2 Scottish Terriers, 2 Cocker Spaniels, 2 Airedales, and 2 Schnauzers. The remaining 35% (23/ 65) consisted of single examples of other breeds. Grossly, the spleens in this study varied considerably in size, color, and form but nonetheless remained distinctive. Marked splenomegaly ranged from 16 to 50 cm (mean 32 cm) in length and varied in proportion to the body size of the dog. The spleens were typically pale red with embedded, contiguous, indistinct nodules producing a smooth, coarsely undulating capsular contour (61%, 20/33) (Fig. 1). The parenchyma was often stippled by distinct but small, irregular foci of pallor (pale yellow) visible through a thin, transparent capsule and on cut surfaces. Distinct, widely separated nodules were evident in a few of the

3 Vet Pathol 36:6, 1999 Splenic Myeloid Metaplasia 585 spleens (Fig. 2). In 43% (15/35) of spleens, there was uniform splenomegaly in which there was no underlying nodularity. The cut surfaces of all spleens bulged, failed to exude appreciable blood, and had a bosselated, tacky surface texture that was coarsely granular. Splenomegaly in these cases was accompanied by increased fragility; i.e., the capsule was easily penetrated and the parenchyma readily torn when handled. Gross (and microscopic) thrombosis and/or foci of splenic infarction (usually apparent grossly) was detected in 69% (45/65) of these cases (Fig. 3). Microscopically, the distinguishing feature of these spleens was extramedullary hematopoiesis (EMH). Hematopoietic cells were often responsible for distortion of the microscopic architecture of the splenic pulp (Fig. 4). Erythropoiesis predominated but occasional myelopoietic foci were evident. In addition to EMH, there was consistent development of discrete or confluent proliferative nodular histiocytic foci, creating additional splenic pulp distortion (Figs. 4 6). These histiocytic aggregates varied from microscopic foci to grossly evident confluent nodularity of splenic parenchyma. In most cases, the histiocytic foci were discrete and contained hematopoietic elements. The larger and more actively proliferative sites sometimes excluded hematopoietic elements and consisted of only pleomorphic histiocytoid components. In some cases, the histiocytic cells had prominent and abundant physaliferous cytoplasm. Erythrophagocytosis was present in 92% of cases (60/65), and, as a result, hemosiderin was consistently present within macrophages, distributed in an irregular pattern throughout the splenic pulp. Hyperplastic periarterial lymphoid sheaths (PALSs) and periarterial macrophage sheaths (PAMSs) were consistantly present in the distorted splenic parenchyma. In some cases, PAMS were an integral part of larger proliferative nodular histiocytic foci (Figs. 5, 6). Multinucleated giant cells were present as a component of the histiocytic proliferation in 23% (15/65) of the spleens (Fig. 7). Ultrastructural features of splenic myeloid metaplasia were characterized by distortion of the lymphoid and sinusoidal relationships of the spleen. A variety of differentiating hematopoietic cells were routinely encountered. Splenic reticular cells and phagocytic macrophages were consistently found in close association and in increased numbers indicative of augmentation in the splenic reticular meshwork (red pulp), causing the lymphoid components to be diminished and distorted (Fig. 8). The cytoplasm of splenic macrophages often contained recognizable erythrocytic fragments or phagolysosomes (Fig. 9). As the splenic histomorphology transitioned to an increasing population of histiocytic cells, fewer hematopoietic cells were present. The ultrastructural characteristics of the cell population included increasing numbers of a transitional cell type with ultrastructural features of both fibroblastic cells and phagocytic macrophages, i.e., fibrohistiocytic cells (Fig. 10). This cytoplasmic architecture was also evident in the multinucleated giant cells encountered as a component of this disease (Fig. 10). Overtly neoplastic cells were histiocytoid in appearance. The cytosol contained abundant, dilated endoplasmic reticulum and varying numbers of lysosomes. Neoplastic histiocytoid cells were regularly observed in direct association with elongated or stellate fibroblastic cells (Fig. 11). With a single exception, in the five examined cases that resulted in the development of systemic lesions (metastasis), the plasma membranes were smooth and uncomplicated. In one of the cases (dog No. 25), plasma membranes were often folded and interdigitated with adjacent histiocytic cells, conforming to the descriptions for dendritic macrophages. 6,9,12 Complete postmortem evaluation of the tissues from 15/65 dogs (23%) demonstrated the presence of metastatic disease in 8/15 (53%). Multicentric organ distribution was as follows: liver, 8/8; bone marrow, 7/7; lymph nodes, 5/8; lungs, 2/8; kidney, 2/8; stomach, 1/ 8; and adrenals, 1/8. Liver tissue from 12 additional dogs was submitted with the spleen or subsequent to splenectomy. Renal tissue from one dog was submitted. The liver tissue showed evidence of histiocytic infiltrates in 7/12 cases (58%). The single renal biopsy submitted with splenic tissue demonstrated diffuse histiocytic infiltrates in the interstitium. Among those dogs subjected to postmortem examination and with no grossly obvious extrasplenic involvement (7/15, 47%), the bone marrow showed an increased cellularity with an underlying histiocytic population in five of seven cases. Although five of seven dogs in this group died spontaneously, and presumably as a direct result of lesions in the spleen (or perhaps the splenectomy itself), an obvious pathologic explanation for death was not apparent. One of the five dogs was profoundly anemic (packed cell volume [PCV], 5%), one showed severe terminal electrolyte imbalance, and the tissue from another suggested the possibility of sepsis and disseminated intravascular coagulation. There were no gross or histopathologic changes in tissues from the other two dogs that could serve as a specific cause of death. Hematology and serum chemistry values that were most often outside the normal range indicated a poorly responsive normochromic, normocytic anemia (reticulocyte median value, 3.5%; PCV median value, 21.9%). Nucleated erythrocytes were consistently noted in the available hematology data. White blood cell (WBC) counts were elevated in approximately 35% (20/58) of the animals (median, /mm 3 ; range

4 586 Spangler and Kass Vet Pathol 36:6, /mm 3 ). Although a consistent pattern of alteration in other routinely evaluated hematology and chemistry parameters was not detected, total bilirubin was elevated in 30% of the dogs (mean, 0.9 mg/dl), and direct bilirubin was elevated in 60% of the dogs (mean, 0.6 mg/dl). The results of Coombs antiglobulin tests were available for 22% of the dogs (14/65). Antibody was detected (at 37 C) in 21% (3/14) and was not detected in 79% (11/14). Antibody was not detected in any of the samples incubated at 4 C. The minimum follow-up interval in this study was 12 months (mean, 24 months; range, months). Only 30% of the dogs survived the followup period; thus, the majority (70%) died during the 12-month postoperative interval. Most of the dead dogs (89%, 47/53) died spontaneously (47%, 25/53) or were euthanatized (42%, 22/53) as a direct result of their splenic disease. While none of the dogs died spontaneously of competing diseases, 8% (4/53) were euthanatized because of conditions unrelated to the splenic disease. The causes of death for 2/53 dogs (4%) could not be accurately determined based on information received at follow-up. The median survival time for all dogs dead from causes related to splenic disease was 1 month (range 0 16 months). Among the distinctive histologic changes present in the splenic tissue and evaluated for prognostic significance (i.e., mitotic activity, erythrophagocytosis, giant cells, thrombosis/infarction, and the proportion and distribution of histiocytic and hematopoietic cells), only the presence of giant cells (hazard ratio, 3.18; P 0.004) was a reliable prognostic feature and indicative of an extremely poor prognosis. All the dogs with splenic giant cells were dead from spleen-related causes at the end of the follow-up period (median time to death, 1 month; range, 0 16 months). None of the dogs living at the end of the follow-up period (12/65, 19%) had shown evidence of giant cells in their spleens. Four dogs (for which there was sufficient extra tissue submitted after death) that did not originally show giant cells in the spleen died with disseminated histiocytic sarcoma lacking giant cells. Discussion The canine spleen is a versatile and dynamic organ. In addition to its well-known functions of blood storage, immune competence, and erythrocyte maintenance, it appears to play a role in some more diverse and less well known processes. This is illustrated by the serendipitous discovery that dogs that had undergone splenectomy were protected from renal circulatory interruption (renal artery occlusion) of up to 120 minutes duration. This is an ischemic time interval that consistently causes severe and sustained renal functional and anatomic degradation in splenic-intact animals, while animals that had undergone splenectomy were largely spared these renal consequences. 7 The fundamentals of veterinary surgical pathology often demand credible recognition and understanding of sometimes complex disease processes based on examination of a single tissue submission. This is especially true of the spleen, for which it is important to recognize lesions that are primary or unique to splenic tissue and those which may involve the spleen secondarily or as a single component of a systemic or multiorgan disease process. Canine splenic myeloid metaplasia/histiocytosis seems to represent a stereotypical response that develops as a consequence of, or perhaps a prelude to, clinical signs of anemia, leukocytosis, and malaise accompanied by splenomegaly. In about 20% (14/65) of the dogs, this splenic lesion was associated with a clinical diagnosis of autoimmune hemolytic anemia. In about 20% of those cases (3/14), immunologic disease was verified by the presence of surface erythrocyte antiglobulins (Coombs test). Based on the pathologic changes in the canine cases of this report 1) splenic myeloid metaplasia, 2) thrombosis and infarction, 3) erythrophagocytosis, 4) splenomegaly, and 5) proliferation of splenic and bone marrow macrophages the following descriptions of recognized human conditions suggest overlapping and perhaps relevant clinical and pathologic features in both dogs and humans. Agnogenic myeloid metaplasia, 19,22 although now considered clonal in humans and thus a neoplastic (myeloproliferative) disorder, is char- Fig. 1. Spleen; dog No. 2, alive 48 months postsplenectomy, myeloid metaplasia/hypersplenism/histiocytosis. Massive splenomegaly (40 cm in length) resulting from multiple poorly defined and confluent nodular masses made up of histiocytic and hematopoietic elements. Fig. 2. Spleen; dog No. 25, dead from related causes 1 month postsplenectomy, myeloid metaplasia/histiocytic sarcoma. Splenomegaly associated with multiple discrete nodular proliferations of histiocytic and hematopoietic cells. Bar 1cm. Fig. 3. Spleen; dog No. 27, dead from related causes 2 months postsplenectomy, myeloid metaplasia/hypersplenism. Splenomegaly associated with diffuse proliferation of histiocytic and hematopoietic elements with segmental infarction (arrows). Bar 1cm.

5 Vet Pathol 36:6, 1999 Splenic Myeloid Metaplasia 587

6 588 Spangler and Kass Vet Pathol 36:6, 1999 acterized by long-term, progressive splenomegaly and trilinear splenic hematopoiesis with cells at all stages of maturation. Splenic infarcts (and presumably thrombi) are a common finding. 22 The rather confusing array of histiocytic disorders/ histiocytosis in human pathology includes both familial (primary) hemophagocytic syndrome and histiocytic sarcomas or malignant histiocytosis. Histiocytosis may involve Langerhans cells or ordinary phagocytic macrophages (histiocytes). Thus, this group of human diseases demonstrates overlapping clinical and pathologic features with the findings for the dogs of the current study. 1,2 Finally, hypersplenism as described in humans is defined by four criteria: 1) cytopenias of one or more peripheral blood cell lines, 2) bone marrow hyperplasia, 3) splenomegaly, and 4) correction of cytopenia following splenectomy. At the present time, however, there is no agreement regarding the specific disorders included in the category, and some disorders for which there is agreement do not fulfill all four criteria. 22 Primary hypersplenism is considered an inherited disorder, while secondary hypersplenism is an acquired condition associated with a wide variety of clinically distinct entities that includes, among others, autoimmune cytopenias, blood parasite diseases, histiocytic disorders, and myeloid metaplasia. 22 All of these conditions produce splenic pathology consistent with canine myeloid metaplasia. The breeds of dogs in this study most often affected by this particular splenic lesion mirrored those of canine splenic disease in general. 15,18 Golden Retrievers were most often involved, followed by Labrador Retrievers and Shepherd-type dogs. As observed previously, 15,16,18 dogs with splenic myeloid metaplasia were significantly younger at diagnosis than dogs with a diagnosis of either benign nodular hyperplasia/hematoma or hemangiosarcoma. 18 Histiocytosis among dogs is a poorly understood phenomenon. One form, cutaneous histiocytoma, although well documented and quite common, has only recently been well defined cytopathologically. 12 Malignant, highly fatal, and disseminated neoplasms consisting of histiocytic cells, often in multiple organs and especially the skin, have been described. 4,8,10,11 The pathologic description of changes in the spleen in canine systemic histiocytosis, although less fatal than other forms of histiocytosis, is strikingly similar to changes observed in some of the dogs in the current study. The hallmark angiocentric orientation of histiocytes was only rarely observed, and cutaneous involvement was not seen in any of the dogs of this study. The lesions reported here are based primarily on surgical submission of the spleen and thus are likely to include a number of clinically distinct syndromes with the common presentation of splenomegaly. While this approach appears to be the basis for useful diagnostic categorization of lesions in canine spleens, it is critically important to integrate the lesions, when possible, with the clinical illness experienced by the patient. Grossly, the changes in these spleens can generally be distinguished as unique but would most often be confused with diffuse infiltrative diseases such as lymphoma. Microscopically, this similarity could be compounded by the existence of pleomorphic and histiocytic variants of lymphoma. However, the patterns of thrombosis, infarction, and myeloid metaplasia seen in the spleens of this study were seldom observed in splenomegaly associated with other splenic diseases. 18 Microscopically, the pattern of histiocytosis, extramedullary hematopoiesis (myeloid metaplasia), and excessive erythrophagocytosis was unique among the spleens examined for this study. 18 In this context, the presence of multinucleated giant cells was reliably predictive of a fatal outcome for those dogs in which giant cells were demonstrated. However, some dogs lacking giant cells also experienced fatal outcomes. Giant cells are a consistent feature of malignant histiocytosis in dogs 4,11 and indeed may be one of the pathologic aspects that defines that disease. Ultrastructural features of the histiocytoid cellular component of these splenic lesions conformed to those previously attributed to histiocytosis in both humans and dogs. 3,4,11,14 Histiocytic cell types were present in Fig. 4. Spleen; dog No. 6, dead from related causes 1 month postsplenectomy, myeloid metaplasia/hypersplenism. A discrete but irregular focus of histiocytic cells (arrows) is present in the red pulp otherwise occupied by hematopoietic elements. HE. Bar 20 m. Fig. 5. Spleen; dog No. 39, alive 32 months postsplenectomy, myeloid metaplasia/hypersplenism/autoimmune hemolytic anemia. The red pulp is distorted by erythroid and myeloid elements, and the periarteriolar macrophage sheath (PAM) is hyperplastic (arrow). HE. Bar 20 m. Fig. 6. Spleen; dog No. 1, dead from related causes 1 month post splenectomy, myeloid metaplasia/hypersplenism. Histiocytic cells predominate in the red pulp isolating hematopoietic cells in small irregular islands. HE. Bar 20 m. Fig. 7. Spleen; dog No. 48, dead from related causes 1 month postsplenectomy, malignant histiocytosis. Pleomorphic histiocytoid cells, with the production of giant cells (arrow), isolate myeloid (m), and erythroid (e) elements in the red pulp. HE. Bar 20 m.

7 Vet Pathol 36:6, 1999 Splenic Myeloid Metaplasia 589

8 590 Spangler and Kass Vet Pathol 36:6, 1999 Fig. 8. Electron micrograph; spleen; dog No. 30, dead from related causes 2.5 months postsplenectomy, myeloid metaplasia/hypersplenism/histiocytosis. A typical area of the splenic parenchyma in which macrophages (M) containing cytoplasmic blood pigment (arrows) are closely associated with splenic reticular cells (R). Lead citrate/uranyl acetate stain. Bar 1 m. Fig. 9. Electron micrograph; spleen; dog No. 30, dead from related causes 2.5 months postsplenectomy, myeloid/ metaplasia/hypersplenism/histiocytosis. Macrophages (M) have engulfed lymphocytes (L) and erythrocytes (E). Lead citrate/ uranyl acetate stain. Bar 2 m. Fig. 10. Electron micrograph; spleen; dog No. 33, euthanatized from related causes 1 month postsplenectomy, malignant histiocytosis. Splenic reticular cell (R) surrounded by basement membrane is closely associated with a number of cells showing cytoplasmic characteristics of both macrophages and fibroblasts (fibrohistiocytic cells [FH]). Portions of a fibroblast (FB) and a multinucleated giant cell (GC) are present. Lead citrate/uranyl acetate stain. Bar 3 m.

9 Vet Pathol 36:6, 1999 Splenic Myeloid Metaplasia 591 Fig. 12. Diagnostic flowchart suggesting the pathologic relationships in canine myeloid metaplasia. the spleens examined in the current study as a continuum ranging from an increase in the number of phagocytic macrophages (sometimes containing blood pigments or fragmented erythrocytes) to pleomorphic histiocytoid cells being present in neoplastic proliferations. In 5 of the 12 cases evaluated by electron microscopy, overt histiocytic neoplasia was indicated by extreme pleomorphism, more intense mitotic activity, and cellular proliferation in nodular areas of the spleen that effectively obliterated the normal architectural features and excluded other cell types. Neoplastic disease was also supported by the knowledge of multiple-organ involvement. The cell types encountered ultrastructurally and their relationships were in many ways, similar to those described previously in splenic nodular lesions designated as fibrohistiocytic prolifer- Fig. 11. Electron micrograph; spleen; dog No. 25, dead from related causes 1 month postsplenectomy, histiocytic sarcoma. Pleomorphic histiocytic cells (H) surround a fibrobast (FB). Lead citrate/uranyl acetate stain. Bar 2 m.

10 592 Spangler and Kass Vet Pathol 36:6, 1999 ation (malignant fibrous histiocytoma). 17 It has recently been suggested, based on the organ distribution of neoplastic foci and cytologic phenotypes, that a relationship exists between the cell types involved in histiocytic proliferative diseases in dogs and those that have been separately classified as malignant fibrous histiocytoma. 5 This conforms with the observations of this study, especially at the ultrastructural level. However, the two conditions may be distinguished at the gross and microscopic levels by the consistent and well-defined nodular nature of malignant fibrous histiocytoma 17 and the presence and prominence of myeloid metaplasia in spleens with coexistent histiocytosis. Postmortem examination and/or examination of additional submitted tissues served to determine the distribution of disease among other organs and provide a cause of death. About half of the cases involved one or more organs in addition to the spleen, suggesting metastasis or, possibly, multicentric origin. The organs involved were among those most often mentioned in reports of canine malignant histiocytosis. 4,8,14 On the other hand, for dogs with single-organ (spleen) involvement, the determination of the cause of death was often hampered by a lack of specific ordefinitive pathologic changes at necropsy. Serum chemistry and hematology values may support a diagnosis of splenic disease but are seldom specific. Most conditions involving splenic lesions can present clinically as anemia with leukocytosis. 13 Mildly elevated bilirubin values in some of the dogs suggested a component of erythrocyte destruction. Positive antiglobulin tests in some dogs confirmed this mechanism acting directly through erythrocyte destruction or erythrophagocytosis; however, splenic phagocytosis of erythrocytes occurred in a large number of dogs in which erythrocyte surface antiglobulins could not be detected. Nucleated erythrocyte response and reticulocytosis, seemingly weak in these dogs, are expected consequences of anemias not resulting from bone marrow dysfunction. The probability of longterm survival following the diagnosis of myeloid metaplasia is extremely poor: 70% mortality in the 12- month follow-up interval. Half of the dogs were dead in the 30-day (1-month) interval following surgery. Death occurred in a pattern similar to that reported previously; 17 i.e., about half of the dogs were euthanatized and half died spontaneously. In the current study, only a few of the deaths were a result of competing illnesses. It should be stressed, however, that 30% of the dogs did survive on a long-term basis ( 12 months). Unfortunately, in the absence of giant cells, no reliable anatomic indicators were evident that would help distinguish dogs destined for longer survival intervals from those with truncated survival. Figure 12 is a flowchart representing the proposed pathologic relationships in canine splenic myeloid metaplasia. Splenomegaly is found to be associated with inappropriate extramedullary hematopoiesis (myeloid metaplasia) in the presence of anemia and erythrocyte phagocytosis. Thrombosis and segmental infarction, although not universal, occur in a majority of cases. These changes, in the presence of an augmented macrophage population, are consistent with functional hypersplenism and/or morphologic histiocytosis. This histiocytosis may be reactive and result from infections or immune-mediated or idiopathic causes, or it may be neoplastic, in which case tumor giant cells are a consistent finding. The relationship between pathologic changes in the spleens from the dogs in the current study and those changes described in the proposed overlapping human conditions of hypersplenism, hemophagocytic syndrome, and agnogenic myeloid metaplasia is apparent. Among human patients, these are clinical syndromes, tightly correlated with corresponding pathologic lesions in the spleen and other tissues. In the canine patients evaluated in this study, the disease consists primarily of descriptive changes in the spleen. Deeper insight into the canine disease(s) will surely require more rigorous integration of the splenic and systemic pathology with the clinical features and response to therapy. In the interim, a diagnostic designation of splenic myeloid metaplasia is a suitably descriptive and preexisting pathologic term 21 for a canine disease that is now more fully described. Acknowledgement We thank Robert S. Clark for invaluable editorial assistance. References 1 Favara BE: Hemophagocytic lymphohistiocytosis: a hemophagocytic syndrome. Semina Diagn Pathol 9:63 74, Favara BE, Feller AC: Contemporary classification of histiocytic disorders. Medi Pediatr Oncol 29: , Franchino C, Reich C, Distenfeld A, Ubriaco A, Knowles DM: A clinicopathologically distinctive primary splenic histiocytic neoplasm, demonstration of its histiocytic derivation by immunophenotypic and molecular genetic analysis. Am J Surg Pathol 12: , Hayden DW, Waters DJ, Burke BA, Manivel JC: Disseminated malignant histiocytosis in a Golden Retriever: clinicopathologic, ultrastructural, and immunohistochemical findings. Vet Pathol 30: , Kerlin RL, Hendrick MJ: Malignant fibrous histiocytoma and malignant histiocytosis in the dog-convergent or divergent phenotypic differentiation. Vet Pathol 33: , Kuehn NF, Gaunt SD: Hypocellular marrow and extra-

11 Vet Pathol 36:6, 1999 Splenic Myeloid Metaplasia 593 medullary hematopoiesis in a dog: hematologic recovery after splenectomy. J Am Vet Med Assoc 188: , Mandal AK, Taylor CA, Bell RD, Hillman NM, Jarnot MD, Cunningham JD, Phillips LG: Erythrocyte deformation in ischemic acute tubular necrosis and amelioration by splenectomy in the dog. Lab Invest 65: , Maxey WL, Davenport DJ, Morton D, Jacobs RM: Malignant histiocytosis in four dogs. J Am Vet Med Assoc 187: , Mierau GW, Favara BE, Brenman JM: Electron microscopy in histiocytosis X. Ultrastruct Pathol 3: , Moore PF: Systemic histiocytosis of Bernese Mountain Dogs. Vet Pathol 21: , Moore PF, Rosin A: Malignant histiocytosis of Bernese Mountain Dogs. Vet Pathol 23:1 10, Moore PF, Schrenzel MD, Affolter VK, Olivry T, Naydan D: Canine cutaneous histiocytoma is an epidermatropic Langerhans cell histiocytosis that expresses CD1 and specific 2 -integrin molecules. Am J Pathol 148: , Neer TM: Clinical approach to splenomegaly in dogs and cats. Comp Cont Educ 18:35 48, Peaston AE, Munn RJ, Madewell BR: Malignant histiocytosis. J Vet Intern Med 7: , Spangler WL, Culbertson MR: Prevalence, type, and importance of splenic diseases in dogs: 1,480 cases ( ). J Am Vet Med Assoc 200: , Spangler WL, Culbertson MR, Kass PH: Primary mesenchymal (nonangiomatous/nonlymphomatous) neoplasms occurring in the canine spleen: anatomic classification, immunohistochemistry, and mitotic activity correlated with patient survival. Vet Pathol 31:37 47, Spangler WL, Kass PH: Pathologic and prognostic characteristics of splenomegaly in dogs due to fibrohistiocytic nodules: 98 cases. Vet Pathol 35: , Spangler WL, Kass PH: Pathologic factors affecting postsplenectomy survival in dogs. J Vet Intern Med 11: , Tefferi A, Silverstein MN: Agnogenic myeloid metaplasia. Cancer Invest 15: , Tubbs RR, Sheibani K, Sebek BA, Savage RA: Malignant histiocytosis. Arch Pathol Lab Med 104:26 29, Valli VEO: The hematopoietic system. In: Pathology of Domestic Animals, ed. Jubb KVF, Kennedy PC, and Palmer N, 3rd ed., vol. 3, pp Academic Press, London, England, Wolf BC, Neiman RS: Disorders of the spleen. In: Major Problems in Pathology, ed. Bennington JL, vol. 20, pp , WB Saunders, Philadelphia, PA, 1989 Request reprints from Dr. W. L. Spangler, 8284 Pleasants Valley Road, Winters, CA (USA). spangler@ solisys.com.