A novel approach for characterization of KSHV-associated multicentric Castleman disease from effusions

Summary

A biopsy of lymphoid tissue is currently required to diagnose Kaposi sarcoma-associated herpesvirus (KSHV)-associated multicentric Castleman disease (KSHV– MCD). Patients showing clinical manifestations of KSHV–MCD but no pathological changes of KSHV–MCD are diagnosed as KSHV inflammatory cytokine syndrome. However, a lymph node biopsy is not always feasible to make the distinction. A pathognomonic feature of lymph nodes in KSHV–MCD is the expansion of KSHV-infected, lambda-restricted but polyclonal plasmablasts. To investigate whether these cells also reside in extra-nodal sites, effusion from 11 patients with KSHV–MCD and 19 with KSHV inflammatory cytokine syndrome was analysed by multiparametric flow cytometry. A distinct, lambda-restricted plasmablastic population (LRP) with highly consistent immunophenotype was detected in effusions in 8/11 patients with KSHV–MCD. The same population was also observed in 7/19 patients with KSHV inflammatory cytokine syndrome. The detection of LRP stratified KSHV inflammatory cytokine syndrome into two clinically distinct subgroups; those with detectable LRP closely resembled KSHV–MCD, showing similar KSHV viral load, comparable severity of thrombocytopenia and hypoalbuminaemia, and similar incidences of hepatosplenomegaly. Collectively, the detection of LRP by flow cytometry can serve as a valuable tool in diagnosing KSHV–MCD. KSHV inflammatory cytokine syndrome with LRP in effusions may represent a liquid-form of KSHV–MCD.

INTRODUCTION

Kaposi sarcoma-associated herpesvirus (KSHV) is the aetiologic agent of Kaposi sarcoma, KSHV-associated multicentric Castleman disease (KSHV–MCD), and primary effusion lymphoma (PEL).¹ Recently, another KSHV-associated condition, the KSHV inflammatory cytokine syndrome, has been described.² Its clinical manifestations and laboratory abnormalities resemble those of KSHV–MCD, but lack the pathological changes diagnostic of KSHV–MCD.

KSHV–MCD occurs predominantly among individuals with human immunodeficiency virus (HIV) co-infection.^3–5 The patients typically manifest periodic episodes of systemic inflammation with lymphadenopathy and constitutional symptoms. Laboratory abnormalities include cytopenia, hypoalbuminaemia, hyponatraemia, and elevated C-reactive protein (CRP). Due to its variable and non-specific manifestations, KSHV–MCD often poses significant diagnostic challenges. Currently, a lymph node biopsy (or examination of splenic specimen) is required to establish the diagnosis.^6,7 The affected nodes typically show increased vascularity and regressed follicles with perifollicular KSHV-infected plasmablasts. These plasmablasts exclusively express IgM and lambda light chain but paradoxically show a polyclonal pattern of immunoglobulin (Ig) gene rearrangement.^6–8 They represent a major source of inflammatory cytokines and are considered the principal pathogenic cells of KSHV–MCD.^9,10 In contrast, KSHV inflammatory cytokine syndrome is a diagnosis of exclusion, requiring clinical exclusion of other inflammatory conditions and pathologic exclusion of KSHV–MCD and PEL. The precise relationship between KSHV inflammatory cytokine syndrome and KSHV–MCD is unclear and the cellular source of cytokine production in the absence of nodal plasmablastic proliferation remains a matter of investigation. Currently, there is no standard therapy for KSHV inflammatory cytokine syndrome.

Pathologic confirmation of KSHV–MCD by lymph node biopsy is limited by multiple factors. First, the histologic features are variable; the presence of concomitant disease such as Kaposi sarcoma in the same lymph node adds an additional layer of complexity. Moreover, a lymph node is not always accessible or a biopsy clinically feasible due to the severe symptoms caused by KSHV–MCD. Importantly, the current pathological criteria for the diagnosis of KSHV–MCD are not applicable in the setting of effusion-based diseases. Persisting or recurring body cavitary effusions are common in patients with HIV and Kaposi sarcoma.¹¹ The aetiology of effusions in these patients can be complex and may represent a manifestation of Kaposi sarcoma itself, KSHV–MCD, PEL, KSHV inflammatory cytokine syndrome, or infection.¹² Accurately distinguishing between these possibilities is difficult but is particularly crucial since they each dictate a substantially different treatment approach. These facts emphasize the need to develop alternative or complementary diagnostic approaches.

Using flow cytometric techniques, we identified a population of KSHV-positive, lambda-restricted plasmablasts in body cavitary effusions of patients diagnosed with KSHV– MCD. The immunophenotype and the monotypic but polyclonal nature of these cells closely resemble those of the KSHV-infected plasmablasts residing in the lymph nodes of KSHV–MCD. Therefore, we conducted a retrospective study including 17 patients with KSHV–MCD and 22 patients with KSHV inflammatory cytokine syndrome but no concurrent PEL or other types of B-cell lymphoma, aimed at evaluating the diagnostic value of detecting this distinct plasmablastic population using multiparameter flow cytometry in diagnosing KSHV–MCD and in determining the aetiology of KSHV-associated effusions.

MATERIALS AND METHODS

Study population

This retrospective study included 39 patients who had either KSHV–MCD (n = 17) or KSHV inflammatory cytokine syndrome (n = 22) but no concurrent PEL or other types of B-cell lymphoma (Table 1 and Supplemental Table S1). They all underwent flow cytometry for evaluation of KSHV-associated diseases in effusions, blood, or bone marrow at the National Cancer Institute (NCI) between January 2012 and January 2022. This study included 62 body fluid specimens from 30 patients and 12 peripheral blood/bone marrow specimens from nine patients who did not present with significant effusions.

TABLE 1.

Laboratory findings and features of patients with body cavity effusions in the setting of KSHV-related diseases

Patient	HIV (copies/ml)	KSHV in effusion (copies/ml)	KSHV in blood (copies/106 cells)	CD4 (cells/μl)	Hgb (g/l)	PLT (K/μl)	Sodium (mmol/l)	Albumin (g/dl)	CRP (mg/l)	IgG (mg/dl)	Kappa LC (mg/dl)	Lambda LC (mg/dl)
MCD with LRP in effusions (n = 8)
1	<50	37 333	2364	240	72	67	142	1.8	27.3	2520	26.2	21.1
2	41 871	1875	33 091	5	81	35	134	1.5	61.9	1330	7.3	8.9
3	79	63 014	3 272 727	113	96	163	130	2.3	NA	1222	16.3	17.5
4	70	97 000	2267	476	89	282	139	3.7	23.7	1583	9.3	12.2
5	<20	0	0	257	91	127	139	3.3	150.5	1958	5.9	4.2
6	132	456 000	857 143	33	84	44	132	1.8	33.3	1031	10.7	25.6
7	NA	NA	1 446 154	238	70	2	137	2.3	101.5	3409	19.0	22.6
8	458	235 294	105 263	84	73	52	139	1.6	216.8	3605	42.4	23.4
Median	132	97 000	33 091	176	80	67	139	1.8	33.3	1770	13.5	19.3
IQR	21 104	326 043	478 888	207	14	134	8	1.2	113.9	1938	16.6	13.5
KICS with LRP in effusions (n = 7)
9	<20	9 538 462	340 000	20	73	29	131	2.0	147	924	6.0	4.0
10	<20	11 084 337	350 000	46	92	81	134	2.4	83.4	1071	6.4	5.4
11	<20	25 455	10 465	198	91	221	134	1.4	65.8	2670	15.2	10.2
12	26	2625	758	511	79	450	131	1.3	13.2	411	5.5	8.8
13	651	161	8 341 674	16	83	116	132	2.3	62.1	1964	15.8	10.3
14	Not detected	7667	3019	31	91	96	128	1.6	31	1070	10.6	12.3
15	<20	147 126	304 000	193	144	86	137	2.9	50.9	962	3.7	2.8
Median	20	77 396	322 000	46	87	91	132	2.2	56.5	1070	6.4	8.8
IQR	167	9 922 921	2 345 464	178	28	132	5	1.0	72.8	1040	9.7	6.3
MCD without LRP in effusions (n = 3)
16	<20	0	94 737	174	94	217	137	2.9	91.3	650	1.3	1.5
17	Not detected	Positive	2204	346	111	263	142	3.2	2.8	745	2.9	1.9
18	83	200	200	167	118	140	138	3.2	91.5	1153	4.0	3.3
Median	52	100	47 469	171	106	179	138	3.1	91.4	902	2.7	2.4
IQR	NA	NA	NA	NA	NA	NA	NA	NA	NA	NA	NA	NA
KICS without LRP in effusions (n = 12)
19	Not detected	538	Positive	115	100	301	134	2.3	29.9	NA	NA	NA
20	123	13 636	0	118	76	154	130	2.8	129.1	3872	119.4	63.5
21	<20	Positive	0	302	101	914	137	3.3	125.4	1325	5.8	3.04
22	1271	5895	1538	82	77	221	131	2.8	71.1	2807	21.2	18.4
23	150 038	0	0	80	88	304	134	3.1	12.9	1950	4.9	3.5
24	35	13 913	343	1	74	138	138	2.5	42.1	2138	11.3	9.3
25	122	349	12 000	25	73	370	140	3.8	9.6	1092	6.8	4.3
26	Not detected	53	48 571	95	87	222	127	2.8	4.6	860	NA	NA
27	Not detected	22	Positive	29	87	311	138	3.2	79.8	877	6.3	3.5
28	45	1053	360	80	81	216	132	3.0	108.7	1816	13.8	7.9
29	Not detected	0	0	NA	100	315	138	2.3	79.7	NA	NA	NA
30	Not detected	0	0	151	101	218	137	3.8	15.8	1555	2.2	3.2
Median	123	3474	352	80	77	219	133	2.9	56.6	2044	12.6	8.6
IQR	38 420	13 444	4154	72	9	171	8	0.6	101.7	1438	39.4	25.6

Abbreviations: CRP, C-reactive protein; Hgb, haemoglobin; HIV, human immunodeficiency virus; Ig, immunoglobulin; IQR, interquartile range; KICS, KSHV inflammatory cytokine syndrome; KSHV, Kaposi sarcoma-associated herpesvirus; LC, light chain; LRP, lambda-restricted plasmablasts; MCD, multicentric Castleman disease; NA, not available/applicable; PLT, platelet.

The diagnosis of KSHV–MCD was made based on the combination of clinical manifestations and pathological confirmation of nodal involvement.⁸ Patients were classified as having KSHV inflammatory cytokine syndrome if they met the previously proposed working criteria (Supplemental Table S2)^13,14 with either a negative biopsy or no definite lymphadenopathy and no PEL. Clinical information was collected from the electronic medical records. Pertinent laboratory findings at the time (+/− seven days) of flow cytometric analysis were obtained, including HIV viral load, KSHV viral load in peripheral blood and effusion, haemoglobin, platelet count, absolute CD4-positive T cells in peripheral blood, CRP, serum immunoglobulin (IgG) and light chains. Additionally, seven patients at initial diagnosis of PEL were also studied to assess whether flow cytometry can reliably distinguish between LRP and PEL. All patients were enrolled on protocols for the study of the natural history of KSHV-related diseases and/or sample collection (NCT01419561, NCT00092222, NCT00006518). These protocols were approved by the NCI Institutional Review Board (and now the NIH Institutional Review Board), registered at ClinicalTrials.gov and conducted in accordance with the Declaration of Helsinki. All patients provided written informed consent. Deidentified participant data are available upon reasonable request made to the corresponding author H-W.W.

Flow cytometry

The specimens were collected and immediately sent to the NCI Flow Cytometry Laboratory for eight-colour flow cytometry at ambient temperature. The preparation of samples included NH₄Cl whole-blood lysis and phosphate-buffered saline wash. Subsequently, the samples were incubated with cocktails of fluorochrome-conjugated antibodies for 30 minutes at room temperature. Intracellular antigen staining (IRF4 and LANA-1) was performed following completion of surface antigen staining, fixation and permeabilization. Cells were then washed, pelleted and fixed in 1% formalin. The specimens were acquired on a FACSCanto II flow cytometer (BD Bioscience) and analysed with FCSExpress software (DeNovo Software). A total of 1 000 000 events were collected per tube when possible.

Viral load studies

KSHV viral load assays were performed using the quantitative real-time polymerase chain reaction (PCR). Peripheral blood mononuclear cells and effusion-associated KSHV viral DNA were quantified using primers from KSHV K6 regions as previously described.² Cell number quantifications were performed using a human endogenous retrovirus 3 (ERV-3) assay.¹⁵ Viral loads were expressed as copies per million cell equivalents and effusion viral loads were expressed as copies/10⁶ mononuclear cells.

Statistical analysis

Student’s t-test or Fisher’s exact test, as appropriate, was performed for comparisons between groups. Analyses were performed with the Statistical Package for Social Sciences (version 17, SPSS Inc.). p-values are 2-tailed; a value of p ≤ 0.05 was considered statistically significant.

RESULTS

Detection of distinct lambda-restricted plasmablasts in effusions from patients with KSHV–MCD by flow cytometry

This study included 17 patients with pathologically confirmed KSHV–MCD and 22 patients with KSHV inflammatory cytokine syndrome; none of them had concurrent PEL or other B-cell lymphomas. Among these 39 patients, 30 (11 with KSHV–MCD and 19 with KSHV inflammatory cytokine syndrome) had developed body cavity effusions that were evaluated by flow cytometry (Table 1).

All 11 patients with KSHV–MCD and effusions were cisgender male, with a median age at presentation of 40 years (range 25–67). All except one (patient no. 17) were HIV co-infected and were receiving antiretroviral therapy; eight (72.7%) had suppressed HIV viral load (<200 copies/ml). All but one (patient no.1) had biopsy-proven Kaposi sarcoma. The median time between HIV, Kaposi sarcoma, and KSHV–MCD diagnosis and first presentation of effusions was 8.5 months (range 0.6–122.2), 0.6 months (range 0–120.2) and 0.3 months (range 0–101.2) respectively.

In eight of the 11 patients (72.7%) with KSHV–MCD, a distinct lambda-restricted plasmablasts population was identified in effusions at least once over the course of the study (Supplemental Table S1). Herein we use LRP to denote specifically this lambda-restricted plasmablast population detected by flow cytometry. The LRP showed high forward and side-scatter light dispersion properties and a distinctive immunophenotypic profile that was highly consistent across all specimens (Figures 1 and 2). It expressed bright CD38, variable CD19, IRF4, Vs38c and CD319 but was dim to negative for CD138 and CD20. This immunophenotype matches the immunophenotypic profile of KSHV-infected plasmablasts seen in lymph nodes of MCD.¹⁶ In the effusions, there was usually a prominent polytypic plasma cell population in the background; LRP can be distinguished by their different antigen expression pattern including higher forward scatter, brighter CD45, dimmer CD38 and VS38c, and dim or absent CD19 and CD138. Importantly, LRP exhibited surface and intracytoplasmic lambda restriction. The number of LRP was variable but typically low, with a median of 0.06% of total mononuclear cells (range, 0.03%–3.1%).

FIGURE 1.

Sequential gating to identify LRP (red). Abbreviations: br, bright; LRP, lambda-restricted plasmablasts; MNC, mononuclear cells

FIGURE 2.

Representative immunophenotype of LRP. Abbreviation: LRP, lambda-restricted plasmablasts. Red, LRP; green, normal plasma cells

To further investigate if LRP represent KSHV-infected plasmablasts, similar to those residing in KSHV–MCD lymph nodes, the expression of latency-associated nuclear antigen-1 (LANA-1) — a KSHV encoded protein that is highly expressed in infected cells — was assessed by flow cytometry in effusions from two patients. The LRP population in both patients demonstrated LANA-1 positivity by flow cytometry (Figure 1). The positivity of LANA-1 was also confirmed by immunoperoxidase staining on cytology preparations of effusions in seven out of eight patients (Figure 3A and Supplemental Table S1). Additionally, six of the eight patients with monotypic LRP demonstrated polyclonal Ig rearrangements in the corresponding specimens by PCR (Supplemental Table S1). These results provided convincing evidence that the LRP in effusions represent the extra-nodal counterpart of the KSHV-infected plasmablasts in lymph nodes. This notion was further strongly supported by the detection of cells sharing the identical immunophenotype using flow cytometry in a lymph node pathologically confirmed with KSHV–MCD (patient no. S2, Supplemental Table S3, Supplemental Figure S1).

FIGURE 3.

Representative cytologic features of KSHV-associated effusion specimens. (A) Atypical lymphoid cells (arrow) with enlarged and irregular nuclei and basophilic cytoplasm were detected in the effusion sample from a patient with KSHV–MCD. Ancillary studies demonstrated positivity of latency-associated nuclear antigen-1 (LANA-1) and lambda light chain restriction in the atypical cells, while in-situ hybridization for EBV (EBER) was negative. (B) Representative effusion specimen from a patient with KSHV inflammatory cytokine syndrome with LRP detected by flow cytometry (KICS + LRP) demonstrated atypical lymphoid cells (arrow) with similar atypical cytomorphology and immunophenotype as those detected in KSHV–MCD samples (A). (C) Representative effusion specimen from a patient with KICS alone demonstrated reactive lymphocytes (left) and histiocytes (right), Abbreviations: EBV, Epstein–Barr virus; KICS, KSHV inflammatory cytokine syndrome; KSHV, Kaposi sarcoma-associated herpesvirus; LRP, lambda-restricted plasmablasts; MCD, multicentric Castleman disease

The use of flow cytometric analysis to distinguish lambda-restricted plasmablasts from primary effusion lymphoma

Primary effusion lymphoma is often the major differential diagnosis when these patients develop effusions. Given that the distinction between these two conditions is often challenging, we assessed whether flow cytometry can reliably distinguish one from the other. To this end, we retrospectively reviewed seven cases of PEL at initial diagnosis. Our analysis revealed that PEL cases displayed considerable individual variability in antigen expression patterns (Supplemental Table S4); importantly, none exhibited an immunophenotypic profile identical to the characteristic phenotype of LRP. Additionally, LRP consistently expressed a restricted lambda light chain, while all but one PEL sample showed absence of cytoplasmic and surface light chain expression (Figure 4, Table 3).

FIGURE 4.

Representative immunophenotype of primary effusion lymphoma. Blue, primary effusion lymphoma

TABLE 3.

Immunophenotypic difference between LRP and PEL

	LRP (n = 19)	PEL (n = 7)	p value
Lambda restriction	19/19	1/7	<0.001
Absent light chain expression	0/19	6/7	<0.001
Bright CD45 expression	19/19	1/7	<0.001
CD81 positivity	17/17	4/6	NS
Background plasmacytosis	19/19	1/7	<0.001

Abbreviations: LRP, lambda-restricted plasmablasts; PEL, primary effusion lymphoma.

A few phenotypic features further distinguished LRP from PEL. Bright CD45 expression comparable to that of mature lymphocytes was a characteristic of LRP, but it was uncommon in PEL. All but one PEL case demonstrated dim to moderate CD45 expression. Furthermore, LRP in effusions were characteristically associated with prominent polytypic plasmacytosis, which was noted in only one PEL specimen. While LRP could be distinguished from normal plasma cells by the differential expression levels of plasma cell markers including CD19, CD138, CD319 and Vs38c, the overall phenotypic profile of LRP still formed a continuum with the plasma cells. In comparison, the immunophenotypic profile of PEL was readily separated from that of normal plasma cells (Figure 4, Table 3). Another helpful feature was the size of the cell populations. PEL clones typically dominated the cellular compartments at initial presentation, comprising a median of 75% of total mononuclear cells (range 24%–97%). In contrast, the numbers of LRP were consistently low in KSHV–MCD or KSHV inflammatory cytokine syndrome. However, this feature may become less reliable in the post-therapy setting.

The presence of lambda-restricted plasmablasts identified a subset of KSHV inflammatory cytokine syndrome closely related to KSHV–MCD

Among the 19 patients diagnosed with KSHV inflammatory cytokine syndrome, LRP were observed in the effusions in nine patients, when present comprising a median of 0.87% (range, 0.01%–8.7%) of the mononuclear cells. Notably, there is a weak correlation between the frequencies of LRP and KSHV viral load in effusion when patients with KSHV–MCD and those with KSHV inflammatory cytokine syndrome were combined for a bivariate correlation analysis (correlation coefficient = 0.67, p = 0.009). Cytologic examination of effusions containing LRP demonstrated scattered large atypical plasmacytoid cells that were morphologically and immunophenotypically similar to those observed in effusions of patients with KSHV–MCD (Figure 3A & 3B). Effusions where LRP were not detected by flow cytometry showed mainly reactive changes only (Figure 3C).

We then queried whether the detection of LRP could serve as a useful marker to illuminate the heterogeneity of KSHV inflammatory cytokine syndrome. We divided KSHV inflammatory cytokine syndrome into two subgroups based on the presence of LRP: KSHV inflammatory cytokine syndrome + LRP (KICS + LRP) versus KSHV inflammatory cytokine syndrome alone (KICS alone). A comparison between the two subgroups revealed several important differences in clinical manifestations and laboratory findings (Figure 5 and Table 2). Patients with KICS + LRP demonstrated more severe hypoalbuminaemia (p = 0.006), a key feature of KSHV–MCD. The presence of LRP was also associated with more evident thrombocytopenia. Five out of seven patients with KICS + LRP had thrombocytopenia (<150 K/μl), which was severe in one (<50 K/μl). In sharp contrast, thrombocytopenia, although considered to be a key diagnostic feature of KSHV inflammatory cytokine syndrome, was observed in only one of 12 patients in the KICS-alone group (p = 0.02). Furthermore, patients with KICS + LRP had more frequent hepatosplenomegaly (p = 0.04).

FIGURE 5.

Comparison of laboratory parameters among patients with different HHV8-related diseases. HHV8, human herpesvirus 8; KICS, KSHV inflammatory cytokine syndrome; KSHV, Kaposi sarcoma-associated herpesvirus; LRP, lambda-restricted plasmablasts detected by flow cytometry; NS, not significant; VL, viral load.

TABLE 2.

Comparison of clinicopathological features among patients with HHV8-related diseases

	MCD	KICS		p value
		With LRP	No LRP	KICS + LRP vs KICS alone	MCD vs KICS + LRP	MCD vs KICS alone
Demographics
Number	11	7	12	—	—	—
Age, yrs	40 (25–67)	42 (24–68)	30 (27–53)	NS	NS	NS
Male/female	11/0	6/1	11/1	NS	NS	NS
HIV/HHV8 parameters
HHV8 in effusion, copies/106 cells	37 333 (0–456 000)	25 455 (161–11 084 337)	701 (0–13 913)	0.032	NS	0.024
HHV8 in blood, copies/106 cells	33 091 (0–3 272 727)	304 000 (758–8 341 674)	352 (0–48 571)	0.023	NS	0.007
HIV, copies/ml	70 (0–41 871)	20 (0–651)	45 (0–150 038)	NS	NS	NS
CD4, cells/μl	174 (5–476)	46 (16–511)	80 (1–302)	NS	NS	NS
Blood cell counts
Hgb, g/l	89 (72–118)	91 (73–144)	81 (73–101)	NS	NS	NS
PLT, K/μl	127 (44–282)	96 (29–450)	221 (138–914)	0.02	NS	0.016
Inflammatory parameters
Albumin, g/dl	2.9 (1.5–3.7)	2.0 (1.3–2.9)	3.0 (2.3–3.8)	0.006	NS	NS
CRP, mg/l	61.9 (2.8–216.8)	62.1 (13.2–147.0)	42.1 (4.6–129.1)	NS	NS	NS
IgG, mg/dl	1330 (650–3605)	1070 (411–2670)	1685 (860–3872)	NS	NS	NS
Kappa LC, mg/dl	7.3 (1.3–42.4)	6.4 (3.7–15.8)	6.8 (2.2–119.4)	NS	NS	NS
Lambda LC, mg/dl	8.9 (1.5–25.6)	8.8 (2.8–12.3)	4.3 (3.0–63.5)	NS	NS	NS
Sodium, mmol/l	139 (132–142)	132 (128–137)	134 (127–140)	NS	0.012	NS
Pathologic features
LRP, % of MNCa	0.06 (0.03–3.10)	0.87 (0.01–8.70)	—	NS	NS	—
Abnormal cytology	9/11 (81.8%)	5/7 (71.4%)	4/12 (33.3%)	NS	NS	0.036
Clonal Ig	3/11 (27.0%)	1/7 (14.3%)	3/12 (25.0%)	NS	NS	NS
Hepatomegaly and/or splenomegaly	6/11 (54.5%)	5/7 (71.4%)	2/12 (16.7%)	0.04	NS	NS
Outcome
Death	6/11	4/7	6/12 (50%)	NS	NS	NS
Complete remissionb	6/11 (50.0%)	3/7 (42.9%)	2/7 (28.6%)	NS	NS	NS
Median survival, mos	86.6	12.1	23.3	NS	NS	NS

Note: For age and laboratory values, figures given are median and range, for other characteristics number and percentage.

Abbreviations: HHV8, human herpesvirus 8; HIV, human immunoideficiency virus; KICS, KSHV inflammatory cytokine syndrome; KSHV, Kaposi sarcoma-associated herpesvirus; LRP, lambda-restricted plasmablasts; mos, months; PEL, primary effusion lymphoma.

^aFrequency of lambda-restricted plasmablasts in the body cavitary fluids, excluding peripheral blood, bone marrow, or peripheral blood.

^bTreatment response in patients who received rituximab-based regimens.

We then set out to assess whether KICS + LRP is more closely related to KSHV–MCD than KICS alone is. Although by definition, KSHV inflammatory cytokine syndrome resembles KSHV–MCD in clinical manifestations and laboratory abnormalities, patients in the KICS-alone group displayed several key parameters distinguishable from those with KSHV–MCD, including a significantly lower KSHV viral load in both effusion and blood (p = 0.02 and 0.007 respectively), higher platelet count (p = 0.016), and seemingly less frequent hepatosplenomegaly (p = 0.08). In contrast, KICS + LRP was similar to KSHV–MCD with respect to most clinical and laboratory features (Figure 5 and Table 2), including age at presentation, levels of haemoglobin, platelets (p = 0.8), albumin (p = 0.2), CRP (p = 0.6) and KSHV viral load in effusion (p = 0.4) and peripheral blood (p = 0.5), frequencies of hepatosplenomegaly and percentages of LRP at presentation. Moreover, patients in both groups showed similar complete remission rates following rituximab-based regimens.

Detection of lambda-restricted plasmablasts in peripheral blood and bone marrow from patients with KSHV–MCD and KSHV inflammatory cytokine syndrome without effusions

To assess whether LRP were also present in other extra-nodal sites, peripheral blood and bone marrow samples were analysed by flow cytometry in patients with KSHV–MCD or KSHV inflammatory cytokine syndrome who did not have effusions. LRP were detected in three of six (50%) patients with KSHV–MCD, and two of three patients with KSHV inflammatory cytokine syndrome (Supplemental Table S3). All five patients with detectable LRP showed thrombocytopenia, and four of them were severe (<50 K/μl). By contrast, thrombocytopenia was seen in two of four patients without detectable LRP in the blood or marrow.

DISCUSSION

By using multiparametric flow cytometry, we demonstrated a body cavitary cellular counterpart of KSHV-infected plasmablasts in the lymph nodes that are considered pathognomonic of KSHV–MCD. Detecting these cells (the LRP) using flow cytometry in extra-nodal sites could serve as a powerful tool in diagnosing KSHV–MCD. Notably, KSHV inflammatory cytokine syndrome could be stratified into two clinically distinct subgroups based on the detection of extra-nodal LRP; those with detectable LRP closely resembled KSHV–MCD. Therefore, we revisit the defining criteria for KSHV–MCD and propose a novel concept of liquid-variant KSHV–MCD.

In this study, LRP were detected in effusions by flow cytometry in eight of 11 (72.7%) patients with pathologically confirmed MCD, and in blood or marrow in three of six (50.0%) patients without effusions. These cells closely resembled the KSHV-infected LRP that proliferate in lymph nodes affected by KSHV–MCD. They expressed plasmacytic markers, including bright CD38 and IRF4; the relatively bright CD45 and dim to negative CD138 indicated that they were at an early stage of plasmacytic differentiation; they exhibited surface and intracellular lambda light chain restriction. The immunophenotype was distinctive, readily distinguishable from normal mature plasma cells, and highly consistent across all patients. Therefore, the gating strategy is straight-forward and consistent; this important advantage ensures the generalizability of this approach to any clinical flow cytometry laboratory.

The detection of LRP by flow cytometry in effusions, blood, or bone marrow can serve as a valuable diagnostic tool in several clinical scenarios. First, it provides a practical and objective tool that supplements lymph node biopsy for evaluation of KSHV-associated diseases especially in patients with HIV. In patients categorized as KSHV inflammatory cytokine syndrome due to lack of lymphadenopathy or absence of pathological confirmation, detection of LRP may identify a subgroup that is closer to KSHV–MCD. In patients having a history of pathologically confirmed KSHV–MCD, this approach is useful in distinguishing MCD effusions from other inflammatory states or other KSHV-associated diseases, especially PEL. The diagnosis of PEL is based primarily on cytologic evaluation, which can be challenging in some cases. The high prevalence of LANA-1-positive LRP in KSHV–MCD cases and prominent cytologic atypia in a subset of these cells highlights that the diagnosis of PEL cannot rely solely on detection of KSHV-positive cells in effusions, e.g. by LANA-1 immunostains. Although clonality analysis can certainly aid in distinguishing between KSHV–MCD and PEL, it is not uncommon to encounter cases that do not follow their typical patterns of Ig gene rearrangements, as has been reported previously and also observed in the present study (Table 2, Supplemental Table S1).^17,18 This study, for the first time, demonstrates that flow cytometry can serve as a valuable adjunctive tool to cytologic examination and clonality analysis in this dilemma. Flow cytometric analysis reliably distinguishes LRP from PEL based on the following features: (1) LRP consistently express restricted lambda light chain, in contrast to the usual absence of light chain expression in PEL. (2) Bright CD45 expression on LRP is a helpful contrast to the dim to moderate CD45 expression in PEL. (3) LRP shows a distinct phenotype of early plasmacytic differentiation and appear to form a continuum with the normal plasma cells usually abundant in the background. Conversely, plasmacytosis is not typically found in PEL and PEL typically exhibits a phenotype readily separate from normal plasma cells. (4) LRP are observed at low frequencies, whereas PEL clones usually dominate the cellular compartments at initial presentation.

Another particularly novel finding of our study is that KSHV inflammatory cytokine syndrome is heterogeneous and can be stratified into two subgroups based on the presence of LRP; this stratification is justified by their different clinical and laboratory findings. Patients in the KSHV inflammatory cytokine syndrome-alone (KICS-alone) group displayed several key parameters distinguishable from those with KSHV–MCD, while the subgroup with detectable LRP (KICS + LRP) is similar to active KSHV–MCD in multiple clinical and laboratory aspects. These findings lead to a pertinent question: is KICS + LRP, in fact, a form of KSHV–MCD despite the absence of nodal involvement? The most widely held belief regarding the pathogenesis of KSHV–MCD is that the systemic manifestations and certain features of the nodal pathology are responses to elevated levels of IL-6 and other circulating factors during the cytokine storm.¹⁹ A major cellular source or driver of these inflammatory cytokines is believed to be KSHV-infected plasmablasts, which have been considered the principal pathogenic cells. Importantly, our past studies have shown that KSHV–MCD and KSHV inflammatory cytokine syndrome show comparable levels of key cytokines, including IL-6, vIL-6 and IL-10, both in the effusions and the serum,^12,14 Based on these notions, the previous question can be posed in a different way: is the compartment where the KSHV-infected plasmablasts proliferate truly a defining attribute of KSHV–MCD, if cells residing in extra-nodal tissues nonetheless result in KSHV lytic replication, pro-inflammatory cytokine deregulation, and eventually the same clinical end-point?

Perhaps, what is most important is a microenvironment to potentiate growth of plasmablasts by autocrine and paracrine effects and cross-infection. Lymph nodes are not necessarily the only compartments that can exert this function. In support of this, marked expansion of the KSHV-infected plasmablasts in bone marrow and spleen, as well as their release to blood, has been observed in patients with MCD.^20–22 The assumption is strongly reinforced in this study, which demonstrates that body cavities could also serve as a reservoir for KSHV replication and a niche for plasmablastic proliferation. Based on the above findings, we propose that KSHV inflammatory cytokine syndrome with an extra-nodal proliferation of KSHV-infected plasmablasts represent a unique liquid-form variant or, alternatively, an early stage (liquid-phase) of KSHV–MCD.²³ Conceptually, there may be a clinical and biological continuum between reactivation of KSHV infection, liquid-form or liquid-phase, and KSHV–MCD with nodal involvement.

In summary, our study highlights the diagnostic utility of LRP detection by flow cytometry in the setting of KSHV-associated diseases. Our observations also reveal that KSHV-associated lymphoid proliferation at extra-nodal sites occurs more commonly than previously appreciated and may contribute to the pro-inflammatory cytokine storm. Finally, we propose that there is a non-nodal form of KSHV– MCD which forms a continuum with nodal KSHV–MCD. A subset of the previously defined KSHV inflammatory cytokine syndrome with detectable LRP falls into this MCD continuum and may benefit from therapeutic regimens for KSHV–MCD.