摘要
Immune checkpoint inhibitor (ICI)-related hemophagocytic lymphohistiocytosis (HLH) (irHLH) is recognized as a rare immune-related adverse event (irAE) with low incidence of 0.03%–0.4% but potentially lethal, with high mortality rate of up to 50% in some cases.1Noseda R. Bertoli R. Muller L. Ceschi A. Haemophagocytic lymphohistiocytosis in patients treated with immune checkpoint inhibitors: analysis of WHO global database of individual case safety reports.J Immunother Cancer. 2019; 7: 117Crossref PubMed Scopus (13) Google Scholar, 2Takeshita M. Anai S. Mishima S. Inoue K. Coincidence of immunotherapy-associated hemophagocytic syndrome and rapid tumor regression.Ann Oncol. 2017; 28: 186-189Abstract Full Text Full Text PDF PubMed Scopus (10) Google Scholar, 3Shah D. Shrestha R. Ramlal R. et al.Pembrolizumab associated hemophagocytic lymphohistiocytosis.Ann Oncol. 2017; 28: 1403Abstract Full Text Full Text PDF PubMed Scopus (25) Google Scholar The main pathophysiologic feature of HLH is an excessive immune cell response with overactivation and expansion of cytotoxic T cells and macrophages, resulting in life-threatening hyperinflammation. These activated immune cells produce large amounts of Th1 proinflammatory cytokines, including interferon-γ (IFN-γ) and IFN-γ-induced chemokines, creating a 'cytokine storm'.4Akashi K. Hayashi S. Gondo H. et al.Involvement of interferon-gamma and macrophage colony-stimulating factor in pathogenesis of haemophagocytic lymphohistiocytosis in adults.Br J Haematol. 1994; 87: 243-250Crossref PubMed Scopus (151) Google Scholar,5Osugi Y. Hara J. Tagawa S. et al.Cytokine production regulating Th1 and Th2 cytokines in hemophagocytic lymphohistiocytosis.Blood. 1997; 89: 4100-4103Crossref PubMed Google Scholar In the absence of validated standards for irHLH, its diagnosis is usually based on the criteria specified in the pediatric HLH-2004 protocol6Henter J.I. Horne A. Arico M. et al.HLH-2004: diagnostic and therapeutic guidelines for hemophagocytic lymphohistiocytosis.Pediatr Blood Cancer. 2007; 48: 124-131Crossref PubMed Scopus (2775) Google Scholar and on the recent development of the HScore.7Fardet L. Galicier L. Lambotte O. et al.Development and validation of the HScore, a score for the diagnosis of reactive hemophagocytic syndrome.Arthritis Rheum. 2014; 66: 2613-2620Crossref PubMed Scopus (443) Google Scholar Here, we report three consecutive patients with metastatic melanoma who developed irHLH while undergoing ICI treatment. The clinical and biological characteristics of the patients as well as the therapy regimens are shown in Table 1. Given the baseline Th1 cytokine profile and the high interleukin (IL)-6 levels in the three patients (supplementary Figure S2A, available at https://doi.org/10.1016/j.annonc.2020.08.2101), tocilizumab (TCZ), an anti-IL-6R antibody, was administered to all three patients, resulting in a rapid improvement of their clinical status with hospital discharge after a median time of 15 days. Similarly, rapid normalization of the biological markers ferritin, fibrinogen, D-dimer and aspartate aminotransferase (ASAT), as well as of cytopenia and coagulation-fibrinolysis disorder, was observed. Tapering of corticosteroid (CS) therapy was terminated at a median of 21 days after TCZ administration (supplementary Figure S1, available at https://doi.org/10.1016/j.annonc.2020.08.2101). The three patients exhibited high baseline levels of circulating proinflammatory cytokines, including IFN-γ and IFN-γ-induced chemokines [IL-6, IL-18, C-X-C motif chemokine ligand (CXCL) 9, CXCL10, CXCL12, C-C motif chemokine ligand (CCL) 5 and IL-1β]. We also observed elevated levels of antiinflammatory cytokines, such as IL-10 and IL-1 receptor antagonist (IL-1RA). The majority of the most important pro- and antiinflammatory cytokines decreased progressively but then normalized with TCZ treatment (supplementary Figure S2A, available at https://doi.org/10.1016/j.annonc.2020.08.2101). Unfortunately, we did not succeed in following the cytokine profile for patient 2 during the early phase after TCZ treatment.Table 1Characteristics, clinical and biochemical data and treatment regimen of the patientsPatient 1Patient 2Patient 3SexMaleMaleMaleAge42 years36 years32 yearsDiseaseMelanoma stage IVMelanoma stage IVMelanoma stage IVCurrent ICIIpi + NivoNivoIpi + NivoPrevious treatmentNoneIpi + Nivo (3 cycles)NoneSymptoms at disease onsetDeterioration of general condition, fever, nausea, extreme fatigue and drowsinessDeterioration of general condition, fever and extreme fatigueDeterioration of general condition, fever and extreme fatigueAdditional work-upFerritin 705.920 ng/ml (N: 30–400)Fibrinogen 0.7 (N: 2–4 g/L)Triglyceride 335.4 mg/dl (N: 0–79)D-dimer >35.200 mg/ml (N < 500)LDH 6.249 U/L (N: 135–225)Haptoglobin <0.10 g/L (N < 0.3–2)Creatinine 399 μmol/L (N: 62–106)Leucocytes 2.5 G/L (N: 4–10)Hemoglobin 7.1 g/dl (N: 13–17) Thrombocytes 124 G/L (N: 150–350)ALT 192 U/L (N: 11–60)CRP 205 mg/L (<10)Ferritin 83.786 ng/ml (30–400)Fibrinogen 1.6 (2–4 g/L)Triglyceride 132.6 mg/dl (0–79)LDH 1.598 U/L (N < 225)Haptoglobin <0.10 g/L (<0.3–2)Hemoglobin 12.8 g/dl (N: 13–17) Thrombocytes 62 G/L (N: 150–350)ALT 78 U/L (N: 11–60)CRP 199 mg/L (<10)ESR 50 mm/h (<10)Ferritin 98.747 ng/ml (30–400)Fibrinogen 0.6 (2–4 g/L)Triglyceride 105.3 4 mg/dl (0–79)D-dimer > 35.200 mg/ml (N < 500)LDH 3.106 U/L (135–225)Haptoglobin g/L < 0.10 (<0.3–2)Leucocytes 3.4 G/L (N: 4–10)Hemoglobin 11.6 g/dl (N: 13–17) Thrombocytes 91 G/L (N: 150–350)ALT 422 U/L (N: 11–60)CRP 44 mg/L (<10) (under CS)Infectious panel negative for:CMV, EBV, VZV, HIV 1/2, HTLV 1/2, HAV, HBV, HCV, HEV, Shiga toxins (Stx1, Stx2), TB spotInfectious panel negative for:CMV, EBV, B19, HIV 1/2, HTLV 1/2, HAV, HBV, HCV, HEV, brucella, coxiella burnetii, bartonella henselae, TB spotInfectious panel negative for:CMV, EBV, HIV 1/2, HTLV 1/2, HAV, HBV, HCV, HEV, coxiella burnetii, TB spotAutoimmune panel negative for:ANA, ANCA, APA, complement (C3, C4, CH50), immunosubtraction, PNH,Direct Coombs' test: C3dCytokine panelAutoimmune panel negative for:ANA, ANCA, AIH, complement (C3, C4)Cytokine panelAutoimmune panel negative for:ANA, complement (C3, C4), immunosubtractionCytokine panelCyTOF panelADAMTS13 normal activity >51%G6PD normalPT 40% (80–120)Fact II 51% (80–150)Fact V 115% (70–180)Fact VII 69% (70–180)Fact X 39% (75–170)PT 75% (80–120)PT 48% (80–120)Fact II 61% (80–150)Fact V 37%(70–180)Fact VII 68% (70–180)Fact X 69% (75–170)Hemophagocytosis features on bone marrow biopsyHemophagocytosis features on bone marrow and liver biopsiesImaging featuresBrain MRI: no evidence of metastasis or hemorrhagic sequelaeBody CT scan: splenomegalyBody CT scan: hepatosplenomegalyConcurrent irAEsPancytopeniaHepatitis G3Disorder of coagulation-fibrinolysis systemAKIN IIIPancytopeniaHepatitis G3Disorder of coagulation-fibrinolysis systemPancytopeniaHepatitis G3Disorder of coagulation-fibrinolysis systemHScore319 points(>99% probability of HPS)288 points(>99% probability of HPS)317 points(>99% probability of HPS)Time point of diagnosis of HLHAfter 2 cycles(51 days)After 5 cycles(78 days)After 3 cycles(91 days)Type and duration of CS therapyHigh dose CS 13 daysTapering over 30 daysHigh dose CS 26 daysTapering over 21daysHigh dose CS 19 daysTapering over 21 daysAdditional treatmentTCZ (2 doses)FFPLow dose heparin prophylaxisTCZ (single dose)Low dose heparin prophylaxisTCZ (single dose)FFPLow dose heparin prophylaxisDuration of hospitalization15 days11 days21 daysCancer responseCRPRPRAIH, autoimmune hepatitis; AKIN III, acute kidney injury stage III; ALT, alanine transaminase; ANA, antinuclear antibody; ANCA, antineutrophil cytoplasmic antibodies; APA, phospholipid antibodies; B19, parvovirus 19; CMV, Cytomegalovirus; CR, complete response; CRP, C-reactive protein; CS, corticosteroid; CyTOF, Mass cytometry; DAT, direct antiglobulin test (direct Coombs' test or direct antiglobulin test); EBV, Epstein-Barr virus; ESR, erythrocyte sedimentation rate; FFP, fresh frozen plasma; HAV, hepatits A virus; HBV, hepatitis B virus; HCV, hepatitis C virus; HEV, hepatits E virus; HIV, human immunodeficiency viruses; HPS, hemophagocytic syndrome; HTLV 1/2, human T lymphotropic virus; Ipi, ipilimumab; LDH, Lactate dehydrogenase; Nivo, nivolumab; PNH, paroxysmal nocturnal hemoglobinuria; PR, partial response; PT, prothrombin time; TB spot, tuberculosis spot; TCZ, tocilizumab; VZV, varicella-zoster virus. Open table in a new tab AIH, autoimmune hepatitis; AKIN III, acute kidney injury stage III; ALT, alanine transaminase; ANA, antinuclear antibody; ANCA, antineutrophil cytoplasmic antibodies; APA, phospholipid antibodies; B19, parvovirus 19; CMV, Cytomegalovirus; CR, complete response; CRP, C-reactive protein; CS, corticosteroid; CyTOF, Mass cytometry; DAT, direct antiglobulin test (direct Coombs' test or direct antiglobulin test); EBV, Epstein-Barr virus; ESR, erythrocyte sedimentation rate; FFP, fresh frozen plasma; HAV, hepatits A virus; HBV, hepatitis B virus; HCV, hepatitis C virus; HEV, hepatits E virus; HIV, human immunodeficiency viruses; HPS, hemophagocytic syndrome; HTLV 1/2, human T lymphotropic virus; Ipi, ipilimumab; LDH, Lactate dehydrogenase; Nivo, nivolumab; PNH, paroxysmal nocturnal hemoglobinuria; PR, partial response; PT, prothrombin time; TB spot, tuberculosis spot; TCZ, tocilizumab; VZV, varicella-zoster virus. Functional assessment before TCZ treatment was carried out in one patient (patient 3). The analysis of the patient's peripheral blood mononuclear cells (PBMCs) showed that the majority of memory CD8+ T cells were from the transitional CD8+ T-cell subset and were highly activated, as reflected by the elevated levels of CXCR3 (76% positive), CD38 (66.7%) and human leukocyte antigen (HLA)-DR (29.2%). Memory CD4+ T cells mainly comprised central memory cells with coexpression of CXCR3 (68%), indicating a strong polarization towards Th1 or Th1/Th17 cells. CD4+ T cells were less activated than CD8+ T cells, as indicated by the lower expression of both CD38 and HLA-DR. The levels of PD-1 and CD25 were low in both CD8+ and CD4+ T cells, suggesting a decrease in the regulatory T-cell population as an effect of ICI treatment. Regarding monocytes, the patient had elevated levels of CD16+ inflammatory monocytes and very low levels of HLA-DR indicative of chronic inflammation. In contrast to HLA-DR expression, CD38 and CD11c expression was within the expected range (supplementary Figure S2B, available at https://doi.org/10.1016/j.annonc.2020.08.2101). The diagnosis of irHLH was also supported by histopathological findings of the bone marrow biopsy (supplementary Figure S2C, available at https://doi.org/10.1016/j.annonc.2020.08.2101) for patient 1 (supplementary Figure S2C,a-d, available at https://doi.org/10.1016/j.annonc.2020.08.2101) and for patient 3 (supplementary Figure S2C,e-f, available at https://doi.org/10.1016/j.annonc.2020.08.2101), both of who showed an increased number of macrophages with engulfed hematopoietic precursors reflecting hemophagocytosis. The liver biopsy of patient 3 also showed similar features of hemophagocytosis (supplementary Figure S2C,g-h, available at https://doi.org/10.1016/j.annonc.2020.08.2101). Given the lack of established therapies, we investigated the possibility of personalized treatment of irHLH according to the patients' cytokine profile. Our patients showed the same dominant signature of high levels of circulating proinflammatory cytokines results observed in reactive HLH forms such as IFN-γ, IFN-γ-induced chemokines (IL-6, IL-18, CXCL9, CXCL10, CXCL12, CCL5 and IL-1β), but not of tumor necrosis factor α (TNF-α),8Billiau A.D. Roskams T. Van Damme-Lombaerts R. et al.Macrophage activation syndrome: characteristic findings on liver biopsy illustrating the key role of activated, IFN-gamma-producing lymphocytes and IL-6- and TNF-alpha-producing macrophages.Blood. 2005; 105: 1648-1651Crossref PubMed Scopus (218) Google Scholar and natural cytokine inhibitors such as IL-1RA9Henter J.I. Andersson B. Elinder G. et al.Elevated circulating levels of interleukin-1 receptor antagonist but not IL-1 agonists in hemophagocytic lymphohistiocytosis.Med Pediatr Oncol. 1996; 27: 21-25Crossref PubMed Scopus (41) Google Scholar and IL-10.5Osugi Y. Hara J. Tagawa S. et al.Cytokine production regulating Th1 and Th2 cytokines in hemophagocytic lymphohistiocytosis.Blood. 1997; 89: 4100-4103Crossref PubMed Google Scholar The concordance of our findings with previous reports suggests a possible shared pathogenesis between irHLH and other reactive HLH forms. Both are characterized by strong Th1 cell polarization, overactivation of cytotoxic T cells, and high levels of CXCL9 and Th1 cytokines. Similar to reactive HLH,10Bracaglia C. de Graaf K. Pires Marafon D. et al.Elevated circulating levels of interferon-gamma and interferon-gamma-induced chemokines characterise patients with macrophage activation syndrome complicating systemic juvenile idiopathic arthritis.Ann Rheum Dis. 2017; 76: 166-172Crossref PubMed Scopus (117) Google Scholar,11Marucci G. Caiello I. Pardeo M. et al.Hemophagocytic lymphohistiocytosis (HLH) mimickers: CXCL9 as a potential biomarker distinguishing HLH from other hyperferritinemic syndromes.Arthritis Rheum. 2018; 70Google Scholar high CXCL9 and Th1 cytokine levels appear to be a potential biomarker for irHLH. However, an interesting and possibly a distinct feature for irHLH was the prominent infiltration of activated CD8+ T cells on the bone marrow biopsy in patient 1. Of note, our patients also showed elevated levels of antiinflammatory regulatory cytokines such as IL-1RA and IL-10. Inhibition of IL-6 signaling represents an attractive target, as IL-6 is an upstream pleotropic proinflammatory cytokine produced during the early stages of inflammation that plays a central role in driving the acute phase of HLH along with the proliferation, survival and activation of cytotoxic T cells.12Li B. Jones L.L. Geiger T.L. IL-6 promotes T cell proliferation and expansion under inflammatory conditions in association with low-level RORgammat expression.J Immunol. 2018; 201: 2934-2946Crossref PubMed Scopus (19) Google Scholar Furthermore, IL-6 signaling amplifies the inflammatory response and contributes as upstream driver of the cytokine storm.8Billiau A.D. Roskams T. Van Damme-Lombaerts R. et al.Macrophage activation syndrome: characteristic findings on liver biopsy illustrating the key role of activated, IFN-gamma-producing lymphocytes and IL-6- and TNF-alpha-producing macrophages.Blood. 2005; 105: 1648-1651Crossref PubMed Scopus (218) Google Scholar,13Strippoli R. Carvello F. Scianaro R. et al.Amplification of the response to toll-like receptor ligands by prolonged exposure to interleukin-6 in mice: implication for the pathogenesis of macrophage activation syndrome.Arthritis Rheum. 2012; 64: 1680-1688Crossref PubMed Scopus (66) Google Scholar Therefore, in contrast to blocking single cytokines further downstream, which might not have any suspensive impact on the activity of the various other cytokines propagating immune dysregulation and end-organ damage, blocking the IL-6/IL-6R axis affects the production of multiple cytokines that contribute to the cytokine storm and inhibits signaling downstream of IFN-γ and several other JAK-dependent cytokines. Hypercytokinemia due to irHLH is an attractive therapeutic target. Suppressing this pathological immune activation with TCZ seems to be a promising strategy for irHLH. In our patients, hypercytokinemia, coagulation disorders and liver and renal impairment quickly normalized after TCZ treatment, allowing for rapid and complete tapering of CS treatment within a median duration of 21 days compared with an 8-week period without TCZ. Hyperferritinemia above 4780 μg/L is a predictive biomarker for death in patients with secondary HLH, with a positive predictive value of 93%.14Grange S. Buchonnet G. Besnier E. et al.The use of ferritin to identify critically ill patients with secondary hemophagocytic lymphohistiocytosis.Crit Care Med. 2016; 44: e1045-e1053Crossref PubMed Scopus (34) Google Scholar Interestingly, even though our patients initially presented with extremely high ferritinemia (between 84 000 and 700 000 μg/L), they survived without any persisting organ dysfunction. Therefore, the occurrence of irHLH is possibly associated with response to ICIs. Some reported cases of irHLH received adjunct etoposide3Shah D. Shrestha R. Ramlal R. et al.Pembrolizumab associated hemophagocytic lymphohistiocytosis.Ann Oncol. 2017; 28: 1403Abstract Full Text Full Text PDF PubMed Scopus (25) Google Scholar or mycophenolate mofetil treatment.15Satzger I. Ivanyi P. Langer F. et al.Treatment-related hemophagocytic lymphohistiocytosis secondary to checkpoint inhibition with nivolumab plus ipilimumab.Eur J Cancer. 2018; 93: 150-153Abstract Full Text Full Text PDF PubMed Scopus (22) Google Scholar Importantly, the duration of CS treatment seemed much longer than that observed in patients also administered TCZ. Moreover, treatment with the anti-IL-6R antibody offers several strategic advantages.16Doms J. Prior J.O. Peters S. Obeid M. Tocilizumab for refractory severe immune checkpoint inhibitor-associated myocarditis.Ann Oncol. 2020; 31: 1273-1275Abstract Full Text Full Text PDF PubMed Scopus (16) Google Scholar, 17Haanen J. Ernstoff M.S. Wang Y. et al.Autoimmune diseases and immune-checkpoint inhibitors for cancer therapy: review of the literature and personalized risk-based prevention strategy.Ann Oncol. 2020; 31: 724-744Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar, 18Kauer J. Horner S. Osburg L. et al.Tocilizumab, but not dexamethasone, prevents CRS without affecting antitumor activity of bispecific antibodies.J Immunother Cancer. 2020; 8: e000621Crossref PubMed Scopus (7) Google Scholar TCZ has the potential to target the effector cells of irHLH without eliciting any significant toxicity. In particular, it is devoid of the myelosuppressive effects of current therapies such as etoposide and can also be administered to patients with impaired renal function, such as patient 1, who had severe acute kidney injury. Taken together, the data regarding the rapid and sustained clinical and biological response to TCZ treatment suggest that dysregulation of the IL-6/JAK/STAT axis contributes to the pathogenesis of irHLH. The results need to be confirmed in a larger patient population. Additional studies may provide insights into the molecular and cellular mechanisms underlying this severe complication of ICI. We thank Prof. C. Sempoux from the Institute of Pathology at the CHUV and Dr Pierre Bohanes, medical oncologist from Clinique de la Source for medical care. Supported by a research fellowship and career development award (to Dr Obeid) from Leenaards Foundation (no grant number).