Idiopathic plasmacytic lymphadenopathy: A conceptual history along with a translation of the original Japanese article published in 1980.

The current consensus on Castleman disease is that it is a group of several distinct lymphoproliferative disorders with different underlying pathogenesis and clinical outcomes. In 1980, Mori et al. proposed the concept of idiopathic plasmacytic lymphadenopathy with polyclonal hyperimmunoglobulinemia (IPL), a disease of unknown etiology, characterized by severe polyclonal hypergammaglobulinemia and generalized superficial lymphadenopathy. After Frizzera et al.'s landmark report in 1983, the term multicentric Castleman disease (MCD) gradually became established, and for a time, IPL was regarded as identical to MCD. However, with the subsequent recognition of human herpesvirus 8 (HHV8)-related MCD in the 1990s and the contributions by Kojima et al. in the 2000s, in which non-HHV8-related MCD (now called idiopathic MCD) was at least subclassified into IPL and others (non-IPL), it is now clear that the original distinctiveness of IPL is still maintained in MCD, which is a diverse collection of diseases.

CASTLEMAN DISEASE

The current consensus on Castleman disease is that it is a group of several distinct lymphoproliferative disorders with different underlying disease pathogenesis and clinical outcomes. According to the lesion distribution, the disorders are classified as unicentric or multicentric. Unicentric Castleman disease (UCD) is further subdivided into hyaline-vascular and plasma cell, types (Figures 1A, 1B, 2, and 3). The hyaline-vascular unicentric subtype, the prototype Castleman disease originally reported by Benjamin Castleman in the early 1950s (Figure 1A),, is now considered a neoplasm of follicular dendritic cells (FDCs), based on the findings that FDC proliferation and dysplastic FDCs can be observed in the lesion, clonal karyotypic and genetic abnormalities are reported, and approximately 7% of FDC sarcomas arise in the background hyaline-vascular Castleman disease. Multicentric Castleman disease (MCD) encompasses a spectrum of disorders with overlapping clinicopathological manifestations, but this name was not widely known until the early 1980s. Although several researchers had reported and suggested a multicentric form of Castleman disease characterized by multiple lymphoid tissue lesions, it had not yet been established as a disease concept.

Fig. 1

Evolution of the concept of Castleman disease

Fig. 2

Histopathology of Castleman disease, hyaline vascular-subtype

Branched vessels with focal hyalinization are observed in the follicle. In this follicle, several regressive germinal centers are grouped by fusion of the thick mantle zones, in which small lymphocytes are concentrically arranged.

Fig. 3

Histopathology of Castleman disease, plasma cell-subtype

The germinal center is well-demarcated from the mantle zone and does not harbor highly branched and hyalinized vessels, clearly different from the vascular changes in the germinal centers of Castleman disease, hyaline vascular-subtype. The mantle zone constitutes concentrically arranged small lymphocytes but is not very thick. Plasmacytosis is observed immediately adjacent to the mantle zone, associated with vascular proliferation.

IDIOPATHIC PLASMACYTIC LYMPHADENOPATHY

In 1978, Shigeo Mori and his mentor, Noboru Mohri, presented a paper with the title “Clinicopathological study of systemic nodal plasmacytosis associated with severe polyclonal hyperimmunoglobulinemia” (translated from Japanese) at the 67th Annual Meeting of the Japanese Society of Pathology (from April 5 to 7, Kumamoto, Japan). They reported four patients with severe polyclonal hypergammaglobulinemia, lymphadenopathy, hepatomegaly, a high erythrocyte sedimentation rate, mild anemia, and hypocholesterolemia. The patients were all male, aged between 10 and 30 years, and there were no obvious triggers for the disease. The initial symptoms included generalized fatigue and mild fever. The serum gammaglobulin and IgG levels were 4 to 6 times higher than normal, and the IgA and IgM levels were also elevated, but M-protein was absent. The disease condition did not markedly worsen, and although the symptoms could be relieved by using steroids, they were not curable. The patients were followed up in outpatient clinics for 3–15 years. No foci of infection or localized bulky lymphadenopathy was found despite a detailed workup. The histological features of the 12 lymph node biopsies were nearly identical. There was a marked increase in the number of plasma cells from the subcapsular region to the deep medulla. The basic lymph node architecture was preserved, but it was difficult to distinguish the cortex from the medulla, and the follicles appeared as islands in a sea of plasma cells. Lymph follicles were composed of a relatively well-developed germinal center and a thin mantle zone and were present from the cortex to the medulla. Vascular proliferation was present in the interfollicular areas but not in the germinal centers. Immunoblasts and hyaline deposits were absent. Paracortical areas were atrophic but present. The plasma cells were a mixture of κ- and λ-positive cells.

In 1980, the term “idiopathic plasmacytic lymphadenopathy with polyclonal hyperimmunoglobulinemia (IPL)” was proposed by Mori et al. as a generic term for an enlarged cohort of 10 cases of unknown etiology characterized by severe polyclonal hypergammaglobulinemia and generalized superficial lymphadenopathy (Figure 1C). Patients with IPL have the clinical and histological features of plasma cell-type Castleman disease.,, Nevertheless, Mori et al. stated that IPL should not be recognized as the same disease entity as plasma cell-type Castleman disease because IPL did not have localized mass formation but presented with generalized lymphadenopathy. Therefore, they did not use the existing terms “giant lymph node hyperplasia,” “Castleman disease,” or “Castleman lymphoma” but coined the term “IPL.” Mori et al. also pointed out that IPL and plasma cell-type Castleman disease may share common pathogenic mechanisms.

MULTICENTRIC CASTLEMAN DISEASE

At the 1980 Annual Meeting of the United States/Canadian branch of the International Academy of Pathology, Glauco Frizzera and his colleagues reported the histopathology and clinical features of the multicentric form in 10 patients. They demonstrated several clinical differences between the localized and multicentric forms. In 1983, they described their morphological observations of the nodal and extranodal involvement of the multicentric form in an enlarged cohort of 15 patients and discussed its etiology and relation to other lymphoproliferative diseases, especially the unicentric form (Figure 1D). After this report, the term MCD gradually became established.

IPL AND MCD

In 1988, Frizzera stated that he borrowed the pathology slides of IPL cases from Mori and concluded on a personal review that IPL and MCD were identical. Correspondingly, in 1991, Mori also stated that the disorder reported by Frizzera et al. was identical to what he and his colleagues reported as IPL. However, considering current knowledge, there appears to be a fundamental difference between the 10 cases of Mori et al. and the 15 cases of Frizzera et al. in terms of histopathologic and clinicopathologic diversity. This difference may be attributed to differences in the case collection criteria. Frizzera et al. collected cases based on histopathologic similarities to UCD, whereas Mori et al. collected cases based on strict clinical criteria, including polyclonal hypergammaglobulinemia, excluding collagen disease and malignancy (Table 1).

Table 1

Criteria for IPL

Original criteria (Mori et al. 1980)

1. Polyclonal hypergammaglobulinemia with serum IgG >4,500 mg/dL, without M-protein2. Generalized superficial lymphadenopathy, the largest of which should be the size of a fingertip on palpation or more than 1.8 cm in the greatest diameter by actual measurement, and a high degree of plasmacytosis on histology, with little or no destruction of the architecture3. No known diseases that could be associated with hyperglobulinemia (infectious diseases, collagen diseases, rheumatoid arthritis and its subtypes, Sjogren syndrome, various allergies including drug allergies, hypersensitivity, so-called adjuvant diseases, myasthenia gravis, hyperthyroidism, hepatitis, liver cirrhosis, and malignant tumors including lymphomas)

Modified criteria (Kojima et al. 2008)

1. Prominent polyclonal hypergammaglobulinemia (γ-globulin >4.0 g/dL or serum IgG level >3,500 mg/dL)2. Multicentric lymphadenopathy3. Absence of definite autoimmune disease4. Normal germinal centers and sheet-like infiltration of polyclonal plasma cells in the lymph node lesion

Non-IPL type was defined by multicentric lymphadenopathy and lymph node lesions containing normal, hyaline vascular (HV), and epithelioid germinal centers as described by Frizzera et al. (1983). Cases demonstrating prominent polyclonal hyperimmunoglobulinemia with HV germinal centers were classified as the non-IPL type.

Abbreviations: IPL, idiopathic plasmacytic lymphadenopathy

The histopathology of the 10 cases of Mori et al. was highly consistent and very similar to that of plasma cell-type UCD reported by Keller et al. The mass of plasma cell-type UCD showed a lymph node structure, with a high degree of plasmacytosis in the interfollicular areas, occasional residual sinuses, normal or slightly hyperplastic germinal centers, and occasional vascularization of the germinal centers, all of which were consistent with the histology of lymph nodes in IPL. IPL was also unique in the prognosis, as most patients had an indolent course but were not completely cured.

In contrast, more histopathologic and clinicopathologic diversity was observed among the 15 MCD cases of Frizzera et al. They subdivided the histological findings of 82 lymph nodes obtained from 15 patients into four types: patterns A, B, C, and D. Pattern C, which was observed in eight lymph nodes (10%) and was characterized primarily by the diffuse, uniform accumulation of plasma cells surrounding secondary follicles, appears similar to the lymph node findings of IPL. As for the prognosis, seven patients had an indolent course characterized by repeated exacerbations and remissions, eight patients had an aggressive course, and nine of the 15 patients died, all but one directly resulting from lymphoproliferative disease or its complications, including two with Kaposi sarcoma and one with lymphoma. According to Frizzera et al.,, patient 14 was a 54-year-old female who presented with fever, weakness, and cough. Her serum IgG level was 5,200 mg/dL. She underwent four lymph node biopsies, all showing pattern C, and was alive 156 months after the onset., Patient 14’s disease appears consistent with IPL, but the remaining 14 patients showed various lymph node findings and clinical symptoms that may or may not be consistent with IPL.

In addition, several clinical and laboratory features of the 15 patients of Frizzera et al. were similar to those of a “collagen disease” to the point that in some of them a diagnosis of rheumatoid arthritis, Sjogren syndrome, or systemic lupus erythematosus was made, although Frizzera described that the overall clinical picture never fitted completely any of these syndromes. In contrast, patients with known diseases associated with hyperglobulinemia, including collagen diseases, were excluded from the study by Mori et al.

These findings indicate that IPL is not identical to MCD but is a more homogeneous disease that is part of MCD, in which characteristic histological and clinicopathological features are shared. Several lines of evidence strongly suggest that interleukin (IL)-6 may mediate these similarities. Frizzera presented the idea that MCD should be designated IL-6 syndrome or, when referring specifically to its pathologic changes, as IL-6 lymphadenopathy, and he subclassified MCD/IL-6 syndrome into primary and secondary forms (Figure 1E).

HHV8 AND MCD

In the early 1980s, the first case officially identified as acquired immunodeficiency syndrome (AIDS) was reported. Subsequently, the association between MCD, Kaposi sarcoma, and AIDS was noted. In 1995, Soulier et al. identified human herpesvirus 8 (HHV8)-like sequences in 14 of 14 HIV-positive MCD cases and in 7 of 17 (41%) HIV-negative cases. In terms of a putative association with HHV8, two of the 15 cases in the Frizzera et al. series had Kaposi sarcoma, suggesting that cases classified today as HHV8-associated MCD would have been included. In 1999, Mori et al. developed a rabbit polyclonal antibody against recombinant ORF73 protein/LANA, and in 2001, examined the prevalence of HHV8 in 75 cases of MCD in the Japanese population and in seven French patients with MCD. One Japanese patient and two French patients had AIDS. HHV8 was detected only in these three AIDS patients, suggesting that HHV8 was unrelated to most MCD in HIV-negative Japanese patients.

IPL, non-IPL, AND TAFRO SYNDROME

Subsequently, Masaru Kojima “reintroduced” the concept of IPL, which even its discoverer had given up referring to and called by another name with some resignation. In 2004, Kojima et al. examined 16 cases of IPL that were confirmed to be negative for HHV8 by immunohistochemistry and suggested that IPL was distinct from the MCD reported in Western countries. Compared with 46 Western patients with MCD in the literature, 16 Japanese patients with IPL had a significantly better 5-year survival rate. Histologically, IPL is characterized by the sheet-like proliferation of plasma cells in the interfollicular area and lymphoid follicles with active germinal centers. In contrast, most reported Western cases of MCD were of the mixed type, with features of hyaline-vascular lymph follicles, sheets of plasma cells, and interfollicular vascular proliferation. The 16 IPL cases reported by Kojima et al. showed neither hyaline-vascular germinal centers nor prominent vascular proliferation in the interfollicular area. In response to this report by Kojima et al., a letter describing the first recognized Western IPL case of a 35-year-old man was published.

In 2008, Kojima et al. examined 28 Japanese patients with idiopathic MCD (Figure 1F). Note that the term idiopathic MCD, first officially used by Kojima in 2008, was defined differently from the same term proposed by Fajgenbaum et al. in 2014 (Figure 1H) and used today, in that the former referred to both HHV8-related and HHV8-negative MCD, whereas the latter indicated only HHV8-negative MCD. Kojima’s idiopathic MCD is a concept identical to Frizzera’s primary MCD (Figure 1E)., However, as mentioned above, HHV8 is positive exclusively in the MCD of HIV-positive patients in Japan; therefore, regardless of either definition, idiopathic MCD almost always refers to HHV8-negative MCD in the MCD of HIV-negative Japanese patients. Among the 28 cases, Kojima et al. identified 18 cases of IPL using their modified criteria (Table 1). In the remaining 10 non-IPL patients, eight were diagnosed as the mixed type and two cases as the hyaline-vascular type. The non-IPL cases exhibited a significant female predominance, higher age distribution, and higher incidence of leukocytosis, thrombocytopenia, pleural effusion, ascites, and autoimmune disease. In 2011, Kojima et al. described IPL as a homogenous disease entity, whereas non-IPL-type MCD was a heterogeneous cluster of disease entities. Among the non-IPL-type MCD cases highlighted by Kojima et al., the disease in a distinct group of cases was later named Castleman-Kojima disease and was found to present TAFRO syndrome, which another group, Takai et al., independently reported in 2010 (Figure 1G).

CONCLUSIVE REMARKS

The original article by Mori et al. was published in The Journal of the Japanese Society for Lymphoreticular Tissue Research in 1980, which was the predecessor of the Journal of Clinical and Experimental Hematopathology. Because it was published in Japanese, the original paper, which would have been epoch-making if it had been published in English, has unfortunately not been widely read worldwide. I am not confident that the authors’ passion, sincerity, and modesty in trying to be as accurate as possible when proposing a new disease concept, which were evident in the original paper, are expressed in my English translation (Supplemental data). The 10 original cases, collected according to strict clinical criteria (Table 1), were highly homogeneous. Therefore, IPL maintains its distinctiveness as an independent entity, corresponding to plasma cell-type idiopathic MCD, not otherwise specified, in the diverse disease concepts of MCD today (Figure 1H), and even in the evolving concept (Figure 1I). Kojima, who passed away at 62 in 2018, would have felt the same way and hoped to keep the concept of IPL from fading away so that researchers worldwide could revisit it.