CD30 as a Target Molecule in the Diagnosis and Therapy of Lymphomas.

1
Introduction
The dominant and strong expression of the TNF Receptor superfamily 8 receptor CD30 by the tumor cells of many lymphoma entities and the minor and low CD30 expression by normal cells in lymphoid tissues has made this molecule an important target for the diagnosis and treatment of lymphomas expressing CD30.
The introduction of CD30 into immunohistochemical diagnostic pathology has led to the definition of anaplastic large cell lymphoma (ALCL) as an entity, to a more precise subtyping of peripheral T‐cell lymphoma (PTCL), to the elucidation of the cellular origin of Hodgkin and Reed‐Sternberg (HRS) cells, and to the understanding of their special features. Moreover, it allowed a clearer differential diagnosis between classic Hodgkin lymphoma (cHL), ALCL with their subtypes, and mediastinal gray zone lymphoma. The precise detection of CD30 in routine biopsies has significantly contributed to the identification of lymphomas suitable for anti‐CD30 immunotherapy.
In the first part of this review, we describe the road to the discovery of the CD30 molecule and the way CD30 has contributed to more precise diagnosis and classification of lymphomas. In the second part of the review, we address how anti‐CD30 therapy was developed and the impact of the anti‐CD30–auristatin conjugate and anti‐CD30 CAR‐T cells in the therapy of CD30‐expressing lymphomas.

2
The Discovery of CD30
2.1
Identification of Many Large Cell Lymphomas as IgM‐Positive B‐Cell Lymhomas (1972)
In the 1960's, large cell neoplasms of lymph nodes were regarded as being derived from reticulum cells or from histiocytes by lymphoma experts in the USA and Europe [1, 2]. This view was mainly based on the dogma that lymphocytes do not have the capacity to transform into large cells. In 1965, immunological studies disclosed that lymphocytes can transform into large cells and consist of B‐ and T‐cells [3]. These discoveries were ignored by the lymphoma experts at that time.
In view of the observation that murine B cells carry on their surface immunoglobulin [4] one of the authors of this review (H.S.) measured IgM in fresh autopsy samples of reticulosarcomas and histiocytic lymphomas. Notably, high amounts of IgM were detected in these tumors, indicating that they were derived from B cells and represented large B‐cell lymphomas of high‐grade malignancy [5, 6].
Subsequent studies aimed at the identification of the cell of origin of the IgM‐negative large cell lymphomas, including the IgM‐negative HRS cells of cHL, were unsuccessful for a long time.

2.2
Establishment of the First Hodgkin Cell Line (1979)
Progress occurred when cell lines of HRS cells were established. However, the first attempts led to false and fraudulent results published in 1977 [7, 8, 9]. This was disclosed 3 years later [10]. In 1979, Diehl's team succeeded in establishing the permanently growing cell line L428 from the pleural effusion of a patient with Hodgkin's disease [11]. It was first shown that the L428 cells resemble HRS cells by the lack of B‐cell, T‐cell, and histiocyte antigens. The subsequent search for a specific or characteristic marker for HRS cells was prompted by the observation that in vivo HRS cells are often surrounded by lymphocytes (Figure 1A). Co‐cultivation experiments of the L428 cells with peripheral lymphocytes from one of the authors (H.S.) revealed that the L428 cells bind lymphocytes to their surface as HRS cells do in vivo (Figure 1B). These findings strongly suggested that the L428 represents a true HRS cell line.

2.3
Search for Hodgkin Related Antigens (1981)
Vaccinations of several rabbits with a highly purified cytoplasmic protein fraction from the L428 cells were performed. The antiserum of one of the rabbits stained selectively the cytoplasm of HRS cells, in frozen sections of a Hodgkin biopsy, following absorption with neutrophils and Daudi cell line cells [12]. These findings suggested that the polyclonal antiserum detects a Hodgkin‐specific molecule, perhaps of viral nature.

2.4
Discovery of the Ki‐1 Antigen (1982)
In view of the above‐mentioned possibility, the monoclonal antibody (mAb) technology—established by Cesar Milstein and George Köhler—was applied to generate a hybridoma producing an antibody of identical quality and unlimited quantity directed at the antigen detected by the polyclonal antiserum. Among the 1500 generated hybridomas, one hybridoma was identified whose secreted antibodies stained in frozen sections the cytoplasm of HRS cells strongly (Figure 1C) and in frozen sections of a human tonsil, few large cells around B‐cell follicles (Figure 1D) [13, 14]. The identified hybridoma was called Ki‐1 [13, 14].

2.5
Classification of Ki‐1 According to the Cluster of Differentiation (CD) as CD30 (1986)
To get the Ki‐1 antibody included in the CD classification it was necessary to have further antibodies with Ki‐1 specificity. In Berlin, eight new Ki‐1‐like hybridomas were generated. The antibodies of these new hybridomas were accepted by the Leucocyte Typing Workshop in Oxford 1986. The Ki‐1 hybridomas were included in the CD cluster classification under the designation CD30 [15].

2.6
Generation of the Ber‐H2 Antibody for the Detection of the Ki‐1/CD30 Molecule in Routine Paraffin Sections (1989)
Among the eight new Ki‐1 hybridomas, one hybridoma with the designation “Ber‐H2” was identified, whose secreted antibodies detected a formol‐resistant epitope of the Ki‐1/CD30 molecule and thus enabled a strong and specific staining of CD30 in routine paraffin sections [16]. With the world‐wide availability of the Ber‐H2 antibody, CD30 was introduced into the diagnostic immunohistologic tissue programs [17, 18].

2.7
Classification of the IgM‐Negative Large Cell Lymphomas as Ki‐1/CD30 Lymphomas (1994)
Surprisingly, the Ki‐1 antibodies, besides HRS cells, also stained the IgM‐negative large cell lymphoma with anaplastic morphology [19, 20, 21], leading to the initial designation of these tumors as Ki‐1 lymphomas. The new anti‐CD30 antibody Ber‐H2 enabled the extension of the investigations about Ki‐1 lymphomas in routine paraffin sections (Figure 2A,B). The initial histologic classification of Ki‐1 lymphomas varied widely and included: malignant histiocytosis, anaplastic carcinoma, pleomorphic large cell lymphoma, and Hodgkin sarcoma. When these Ki‐1 positive cases were reviewed together, they appeared to belong to a single entity, designed as ALCL. This was accepted by the R.E.A.L [17] and WHO lymphoma classifications [18]. Subsequent immunoglobulin and T‐cell receptor (TCR) rearrangement studies identified Ki‐1/CD30‐positive tumors as a special T‐cell derived lymphoma entity with frequent loss of T‐cell program [22].

3
Features of the CD30 Molecule
3.1
Chromosomal Localization and Structure of the CD30 Molecule (1992)
The CD30 gene is localized at chromosome 1p36.13 (Figure 3, left), closely linked to other members of the TNF receptor superfamily, such as the human TNFR2 and OX40 genes [23, 24] and represents a member of the Tumor Necrosis Factor Receptor (TNFR) superfamily [25, 26]. CD30 is a 120 kD glycoprotein with intracellular, transmembrane, and extracellular domains (Figure 3, right). The extracellular domain consists of six cysteine‐rich regions in a duplicated structure [25, 26, 27, 28]. All TNF family proteins, including CD30, form homotrimers, a configuration essential for their functionality. The cytoplasmic end of the CD30 protein contains TNF receptor associated factor binding sequences [27] that can activate the pathway of the nuclear factor kappa B (NFkB) transcription factor [29, 30]. The mature form of CD30 is processed from a precursor of about 84 kD during its passage through the Golgi complex [26]. The molecular‐weight shift from 84 to 120 kD is mostly due to glycosylation [31, 32]. The extracellular part of CD30 is cleaved proteolytically by a zinc metalloprotease (ADAM17). This soluble CD30 is released in the serum where it is detectable both in cHL [33] and ALCL [34] patients. One question is whether the soluble CD30 may bind and neutralize parts of the injected anti‐CD30 immunotoxin. However, reduction of therapeutic effect by the soluble CD30 appears unlikely since the immunotoxin is administered in excess as compared to the minimal amount of native anti‐CD30 monoclonal antibody that has been shown to target tumor cells in vivo [35].

3.2
Biological Function of the CD30 Protein
Epitope stimulation experiments of CD30 resulted in receptor trimerisation and signal transduction via recruitment of TNFR associated factors (TRAFs) and TRAF‐binding proteins, generating a signaling complex [36, 37, 38]. TRAF2, as well as TRAF1 and TRAF5, are all implicated in the signaling process. Downstream effects of CD30 stimulation are mediated in part by NFkB, as well as by mitogen‐activated protein kinases/extracellular signal‐regulated kinase pathways [29, 38, 39, 40, 41, 42, 43]. A novel domain in the CD30 cytoplasmic tail also mediates NFkB activation, without direct interaction of TRAF2 or 5, suggesting involvement of unknown TRAF protein(s) in the signal transduction pathway of CD30 [44].
The high and consistent expression of CD30 in cHL and ALCL and its rare and low expression in normal lymphoid tissues suggests that CD30 plays a significant role in the development of cHL and ALCL [45]. The finding that the NFkB and mitogen‐activated protein kinase/extracellular signal‐regulated kinase pathways are integral to CD30‐mediated signaling appears to be a hint that CD30 expression may confer a proliferative and anti‐apoptotic benefit in neoplastic cells [29, 45, 46]. Horie et al. [43] proposed a link between CD30 overexpression and ligand‐independent stimulation of the NFkB pathways in cHL cells, underscoring a possible link between CD30 expression and tumor perpetuation. The group of Stein et al. [45] could not replicate these findings and suggested that NFkB activation in cHL is constitutive and unrelated to CD30 but present in ALCL cells. Watanabe et al. [47] showed CD30 upregulation in cHL and ALCL cell lines might be linked by a self‐perpetuating loop through the mitogen‐activated protein kinase/extracellular signal‐regulated kinase pathway to the expression of JunB, a member of the activator protein (AP‐1) transcription factor family, with diverse effects including a possible link to malignant transformation. It was shown in ALCL cell lines that the transcription factor interferon regulatory factor 4 (IRF4) drives CD30 expression in a positive feedback loop involving NFkB [48]. In addition to IRF4 and AP‐1/JunB, the Ets transcription family has been implicated in tumor cell upregulation of CD30 [48, 49, 50, 51].
Additional studies sought to define the role of CD30 stimulation in lymphoma pathogenesis but were hampered by the application of differing ligands and different cell lines. The interpretation of the different studies and the pleiotropic effects of CD30 stimulation resulted in a remarkable degree of controversy. Thus, it is ultimately unclear whether CD30 contributes to the pathogenesis of CD30 expressing lymphomas.

4
Features of CD30‐Positive HRS Cells of cHL
4.1
Frequency of HRS Cells
When first described, the frequency of HRS cells assessed in hematoxylin–eosin stained sections of diseased tissue was 0.5%–2% [52, 53]. Mononuclear Hodgkin cells were often not recognized histologically because their morphology resembled that of other non‐malignant large cells. The low frequency figures stem from the time before the availability of the CD30 immunostaining. The inclusion of CD30 into the diagnostic programs showed that the frequency of HRS cells can reach values of > 10% in Hodgkin tissue biopsies (Stein H, unpublished findings).

4.2
Elucidation of Cellular Origin of HRS Cells by Picking CD30‐Positive Cells From Frozen Sections of Hodgkin Biopsies
The many studies to clarify the cellular origin of HRS cells failed because of the rareness of the HRS cells in the Hodgkin‐affected tissues. CD30 immunostaining made single HRS cells visible for their extraction from frozen sections. The isolated CD30‐positive HRS cells were subjected to single‐cell PCR and proved to contain clonally rearranged IG in the V genes, a hallmark of malignant B‐cells [52, 54, 55, 56]. High load of somatic mutations was found in the clonally rearranged VH genes of HRS cells, suggesting that HRS cells are derived from germinal center B cells [52]. Unlike other B cell antigens, one molecule of the B‐cell program, PAX5, was not switched off. Its expression was a phenotypical confirmation of the B‐cell nature of HRS cells [57]. Based on these findings, the previous designation Hodgkin's disease was changed to Hodgkin lymphoma (cHL) [18].

5
CD30 Expression in Normal and Pathological Tissues
5.1
CD30 Expression in Normal Tissues
In normal lymphoid tissues, CD30 is expressed in a small subset of large mononuclear cells with evident nucleoli [13]. These large CD30‐positive cells are in the cell cycle as evidenced by Ki‐67 expression [58] and are mostly localized around B‐cell follicles and, to a minor degree, at the edge of germinal centers of tonsil, lymph nodes, and spleen, and around the Hassal's corpuscles of thymus [13, 16, 19, 21]. The cytological features of CD30+ cells and their topographical distribution in normal lymphoid tissues recall that of HRS cells in cHL. Therefore, it has been suggested that these elements might represent the normal counterpart of the neoplastic population of cHL [13, 16, 21]. Immunophenotypic and IgV gene analyses demonstrated that the rare germinal center and extrafollicular CD30+ cells represent B cells [59]. The transcriptomes of CD30+ germinal and extrafollicular B cells shared a strong MYC signature that differed from those of the CD30‐negative GC B cells, memory B, and plasma cells. CD30+ germinal center B cells may represent MYC+ centrocytes redifferentiating into centroblasts, whilst CD30+ extrafollicular B cells may consist of active, proliferating memory B cells [59]. By further investigations, it was found that the many CD30‐positive cells in reactive lymph nodes are a heterogenous population of polyclonal B cells, leading to the conclusion that there is no indication that such CD30‐positive B‐cell populations represent precursor lesions of Hodgkin lymphoma [60].

6
The Diagnostic Impact of CD30 Molecule
CD30 expression has been detected in various lymphoid malignancies, the highest and most consistent expression being observed in cHL and ALCL. So the detection of high levels of CD30 is a must for the diagnosis of cHL and ALCL (Figure 4). More recently, the role of flow cytometry detection of CD30 for the diagnosis of breast implant‐associated ALCL has been assessed [61]. It is important to note that the presence of CD30‐positive cells in peri‐implant fluid does not equate to a diagnosis of lymphoma unless atypical cells are present. Variable expression and intensity of CD30 have been found in T‐cell lymphomas, like PTCL, cutaneous T‐cell lymphoma (CTCL), and extranodal NK‐T‐cell lym [62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72].
CD30 expression can also be a hint for the presence of EBV‐infected cells in malignant lymphoproliferations like EBV‐positive mucocutaneous ulcer and EBV‐positive DLBCL, as well as in non‐malignant lymphoproliferations like infectious mononucleosis [71] and chronic EBV‐driven lesions. Frequent CD30 expression also occurs in aggressive systemic mast cell tumors and leukemia, soft tissue tumors with kinase gene fusions, and germ cell tumors, most consistently in embryonal carcinoma [73, 74, 75, 76, 77]. Table 1 shows the lymphoma entities, other tumors, and non‐neoplastic lesions in which CD30 expression has strong, moderate, or only little or no diagnostic importance.

7
Development of Anti‐CD30 Immunotoxin
Because of its strong and consistent expression in HRS and ALCL cells and the limited expression in normal lymphoid tissues, the CD30 molecule was considered a suitable immunotherapeutic target [21, 35, 78].
7.1
Development of First Anti‐CD30 Immunotoxins (1992)
Pre‐clinical studies with the native anti‐CD30 mAbs SGN‐30 and 5F11 had shown promising activity in vitro and in mouse models [79, 80, 81], but only demonstrated scarce anti‐tumor efficacy in patients [82, 83, 84].
We investigated the activity of anti‐CD30 antibodies conjugated with the plant toxin saporin or other toxins preclinically [78, 85, 86] and demonstrated for the first time that this immunotoxin was active in patients with refractory/relapsed cHL [87]. However, relapses occurred frequently because repeated treatments were not possible due to the strong host immune response both against saporin and ricin and the murine antibody moieties [87, 88].

7.2
Development of a Potent Anti‐CD30 Monoclonal Antibody Auristatin Conjugate (2003)
Eleven years later, the non‐immunogenic toxin monomethyl auristatin E (MMAE) was linked to a humanized anti‐CD30 antibody. Auristatin is a microtubule inhibitor [89, 90]. The anti‐CD30 auristatin conjugate binds to the surface CD30 of neoplastic cells and is internalized through receptor endocytosis and captured into lysosomes where the auristatin is released by proteolytic enzymes and becomes toxic. The released auristatin inhibits the polymerization of tubulin in the cellular cytoskeleton, with consequent cell cycle arrest in G2/M and apoptosis or antibody‐dependent cellular phagocytosis. Auristatin can also induce the death of neighboring cells by diffusion across the cell membrane [91, 92]. Moreover, auristatin seems to exert additional anti‐tumor immunity by stimulation of dendritic cells [93]. The conjugate was named brentuximab vedotin (BV) and has shown significant in vitro and in vivo activity in preclinical and clinical studies [89, 94, 95].
Despite its high specificity, BV has off‐target effects, especially in terms of peripheral neuropathy that is likely mediated through the anti‐tubulin effects of auristatin [91]. Other, more rare, but serious side effects include pancreatitis and leukencephalopathy. Various resistance mechanisms to BV have been reported in cHL and ALCL cell lines, including increased expression of drug transporter proteins [96], auristatin E resistance [97] and release by HRS cells of CD30‐containing extracellular vesicles [98]. CD30 downregulation has been observed in vitro but not in tissue samples of cHL [99]. However, downregulation of CD30 following BV has been reported in refractory (relapsed (R/R)) ALCL cases [100, 101, 102].

8
CD30 as a Target in cHL
Outcomes for patients with advanced‐stage cHL have improved dramatically over the past half century. However, up to 30% of patients with stage III–IV cHL show refractory disease or relapse after frontline chemotherapy, and the outcome is even worse in older patients. During the last two decades, further progress has been made using anti‐CD30 immunotherapy.
8.1
Brentuximab Vedotin Plus Chemotherapy Reshapes Frontline Treatment of Advanced Stage cHL
Until recently, ABVD (doxorubicin, bleomycin, vinblastine, and dacarbazine) was the standard of care in cHL in the USA. The German Hodgkin Study Group (GHSG) used mainly escalated BEACOPP (bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone) (esc‐B) for advanced cHL.
A change in the standard of care in this disease was the inclusion of BV in the first‐line treatment [94, 103]. The ECHELON‐1 phase 3 trial compared in stage III–IV cHL, BV‐AVD (n = 664 patients) to ABVD (n = 670 patients) [94, 104]. The 2‐year modified PFS was 82.1% for BV‐AVD and 77.2% for ABVD (p = 0.04) [104]. There was a higher incidence of neutropenia (58% versus 45%) and ≥ 2 degree peripheral neuropathy (31% versus 11%) in BV‐AVD vs. ABVD. Pulmonary toxicity of grade 3 or higher was reported in < 1% of patients receiving BV‐AVD and in 3% of those receiving ABVD. At 6‐year follow‐up, BV‐AVD had a 7.8% PFS and a 4.5% OS benefit over ABVD [105, 106].
BrECADD (BV, etoposide, doxorubicin, cyclophosphamide, dacarbazine, dexamethasone) was compared with escalated BEACOPP (esc‐B) in the PET‐adapted phase 3 HD21 trial [107, 108]. Sixty‐four percent of patients were iPET negative and received a total of 4 cycles of therapy. As compared to esc‐B, BrECADD showed a superior 4‐year PFS (94.3% versus 90.9%), lower acute hematologic toxicity (G4 31% versus 52%), and lower peripheral sensory neuropathy (all grades 38% versus 49%). There was no significant residual organ toxicity at 1 year after treatment, and > 95% of female patients had hormonal recovery [109]. Based on these results, BrECADD is the new standard of care within the GHSG for adult cHL patients with advanced stage disease, age ≤ 60 years.
Older patients with advanced cHL, especially those unfit for chemotherapy, usually have worse outcomes than younger patients [110]. However, BV monotherapy [111, 112] or sequential BV‐AVD [113] has shown promising results in this setting. With sequential BV‐AVD, the 2‐year PFS and OS were 84% and 93%, respectively. Patients with cHL > 60 years unfit for chemotherapy may also benefit from BV given in combination with dacarbazine or nivolumab [114]. The response rates were high, and, with a median follow‐up of > 4 years, nearly one‐half of responses were durable [114]. Long‐term follow‐up with BV plus nivolumab indicates that this regimen is effective for patients > 60 years, with a cure rate of approximately 40% [115]. BV‐AVD was highly active and had a tolerable adverse event rate even in HIV‐related, advanced‐stage cHL [116, 117].

8.2
Reshaping Frontline Treatment of Early‐Stage cHL
BV combinations were evaluated in early‐stage unfavorable cHL with the aim of increasing the CR rates, reducing the duration of therapy, or omitting radiation therapy. In the BREACH study [118], four courses of BV‐AVD followed by involved nodal radiotherapy (30 Gy) were associated with higher PET‐2 negative rates as compared to 4 ABVD + 30 Gy (82.3% vs. 75.4%). However, 2‐year PFS was similar (97.3% versus 92.6%). Febrile neutropenia and peripheral neuropathy were more frequent with BV‐AVD. In another study, patients with newly diagnosed, early‐stage cHL with unfavorable risk were treated with 4 cycles of BV‐AVD, and those achieving negative PET received 30 Gy in situ radiation therapy [119]. More than 90% of patients achieved negative interim PET after 2–4 cycles of BV‐AVD [119]. All 25 patients who completed BV + AVD + in situ radiation therapy achieved a CR. With a median follow‐up of 18.8 months, by intent to treat, the 1‐year PFS was 93.3%. The treatment was well‐tolerated with no significant pulmonary toxicity. This may be a highly active regimen, especially in patients with bulky disease. In a phase 2 study, ABVD (2–6 cycles) followed by consolidation BV (6 cycles) was evaluated in unfavorable risk, early‐stage cHL [120]. The CR rate was 95% with an estimated 3‐year PFS of 92%.
BV‐AVD was also evaluated for non‐bulky stage I/II cHL [121] resulting in a high CR rate and PFS and OS of 94% and 97%, respectively, with most patients requiring only 4 cycles of therapy. Because toxicity was higher than would be expected from AVD alone, this approach may not be appropriate for early‐stage patients with a favorable prognosis. Finally, a study evaluated the omission of bleomycin and vinblastine (BV‐DC) in non‐bulky limited‐stage cHL. A high CR rate and durable PFS were observed, with most patients requiring 4 cycles of therapy [122]. The estimated 5‐year PFS and OS were 91% and 96%, respectively.
Cumulatively, all the above studies suggest that high CR rates and excellent PFS may be obtained by incorporating BV, allowing a chemotherapy‐only treatment of early‐stage cHL.

8.3
Brentuximab Vedotin in Refractory/Relapsed (R/R) cHL
R/R cHL is usually treated with non–cross‐resistant chemotherapy regimens followed by autologous stem cell transplantation (ASCT) in chemosensitive patients. This combination usually results in about 50% long‐term cure rates. BV in R/R cHL was evaluated in two studies, one of phase 1 [123] and the other of phase 2 [124], both showing safety and activity. In the largest study of phase 2, 102 patients with R/R cHL after ASCT were enrolled [125]. The ORR was 75% with 34% CR. The median PFS for all patients was 5.6 months, and the median duration of response for those in CR was 20.5 months. After 1.5 years, 31 patients were alive and free of documented progressive disease. The most common treatment‐related adverse events were peripheral sensory neuropathy, nausea, fatigue, neutropenia, and diarrhea. Five‐year follow‐up data demonstrate that a subset of patients with R/R cHL who obtained CR with single‐agent brentuximab vedotin had long‐term disease control and may potentially be cured [124].
BV as a single, second‐line agent may be an effective bridge to ASCT, with an ORR of 75% and a CR rate of 43% [126]. After ASCT, the 2‐year PFS and OS were 67% and 93%, respectively. Pre‐transplantation PET negativity was one of the strongest predictors of outcome post‐transplant. A multicenter phase 2 study reported an ORR of 68% after BV x 4, and with this approach, 49% of patients (18/37) avoided additional chemotherapy, proceeding directly to ASCT after CR (35%) or partial response (14%) [127]. BV plus nivolumab was also safe and effective as a bridge to ASCT, with an ORR of 85% and a CR rate of 67%. OS at 3 years was 93% [128].
PET‐adapted sequential salvage therapy with BV followed by augmented ICE resulted in a high rate of PET‐negativity, with a 2‐year PFS of 80% post‐transplant [129]. Lynch et al. [130] observed similar outcomes combining dose‐dense BV and ICE. Due to its significant toxicity, this approach may be a second‐line therapy option in younger transplantation‐eligible patients. In addition to ICE, BV was combined with other chemotherapy regimens, resulting in CR rates of 70%–79% and 2‐year PFS rates of 70%–76% after ASCT [131, 132, 133]. However, these regimens are associated with a high rate of grade 3–4 hematological toxicity. BV with or without salvage chemotherapy appears to enhance PFS in cHL patients with relapsed disease but not in those who are primary refractory [134].
BV as consolidation after ASCT represents a treatment option for cHL patients with unfavorable risk factors (refractory disease, relapse < 1 year after initial treatment, extranodal disease at relapse) who usually progress after ASCT [135]. The randomized, double‐blind, phase 3 AETHERA trial evaluated the safety and efficacy of BV as consolidation in this setting [136]. In this study, 329 patients with R/R unfavorable‐risk cHL were randomly assigned to receive either BV or placebo for up to 16 cycles after ASCT. The median PFS in the BV and placebo groups was 42.9 and 24.1 months, respectively [136]. However, no significant difference in OS was observed between BV and placebo arms [136]. At a further follow‐up, BV continued to provide patients with sustained PFS benefit, 5‐year PFS being 59% with BV vs. 41% with placebo [137]. These results led to the FDA approval of BV as post‐ASCT consolidation in cHL patients at high risk of relapse or progression.
Cumulatively, the above studies suggest that BV can be combined safely with chemotherapy, resulting in high CR rates and a potential to improve outcomes with ASCT [105].

9
Targeted Therapy of CD30+ Peripheral T‐Cell Lymphomas
9.1
Reshaping Frontline Treatment of CD30+ ALCL and Other PTCL
CHOP represents the standard front‐line therapy for patients with PTCL, but this regimen is associated with frequent relapses and a 5‐year survival of only 35% [138, 139]. To improve clinical results, BV has been combined with chemotherapy in the frontline setting [140, 141, 142]. The phase III ECHELON‐2 compared BV + cyclophosphamide, doxorubicin, and prednisone (CHP) vs. CHOP in patients with ALCL or other CD30‐positive PTCL with > 10% CD30 expression [143]. At the 5‐year follow‐up, BV‐CHP was superior to CHOP, with improved PFS (51.4% vs. 43%) and OS (70.1% vs. 61%) [144]. A further analysis of ECHELON‐2 showed that improved outcomes at 5 years of follow‐up were observed with A + CHP vs. CHOP in both ALK+ and ALK−subgroups [145].
BV‐CHP is clearly the standard of care option for ALCL that show strong expression for CD30 whilst the benefit of this regimen in PTCL‐NOS and AITL subgroups that express CD30 not uniformly is uncertain. The impact of consolidative ASCT after BV + CHP in patients achieving CR at the end of treatment was assessed in a subgroup analysis of the ECHELON‐2 trial, which reported better PFS with ASCT than without (5‐year PFS 65.3% vs. 46.4%) [144]. These results were also confirmed in another study [146].
CHOEP has higher activity than CHOP, especially in ALK+ ALCL [139]. The safety and efficacy of CHEP‐BV followed by BV consolidation have been evaluated in 46 patients with CD30‐expressing PTCL. ORR was 91% and CR was 80%, regardless of subtype or CD30 expression [147]. This regimen may have a role in young, high‐risk patients. These data were confirmed in a more recent study [148]. CHEP‐BV was safe and effective even as frontline therapy in adult T cell lymphoma/leukemia, when used as a bridge to allogeneic hematopoietic stem cell transplantation [149].
BV in CD30+ cutaneous T‐cell lymphoma and lymphomatoid papulosis is well tolerated and results in an ORR of 73% and a CR of 35% [150]. The activity and safety of BV versus physician's choice of either methotrexate or bexarotene were evaluated in cutaneous T cell lymphomas in a phase III trial [151]. Significant improvement in objective response lasting at least 4 months was seen with BV vs. physician's choice of methotrexate or bexarotene (56.3% vs. 12.5%) [151].

9.2
BV in R/R CD30+ ALCL and Other PTCL
About half of CD30+ ALCL patients, especially the ALK‐negative form and an even higher percentage of other CD30+ PTCL subtypes, relapse after front‐line treatment and consolidation with ASCT [152, 153, 154]. Therefore, BV has been widely used in the R/R setting. The safety and efficacy of BV as monotherapy (1.8 mg/kg, once every 3 weeks for up to 16 cycles) were evaluated in a Phase II study in 58 patients (median age 52 years) with R/R ALCL, mostly ALK‐negative [155]. The ORR was 86% in patients who had failed prior therapies including ASCT, with 57% CR. The median duration of the response was 13.2 months for CR. Neutropenia, thrombocytopenia, and peripheral sensory neuropathy were the principal grade 3–4 adverse events affecting > 10% of patients. At a 5‐year follow‐up, 38 (66%) patients were in CR, independently of ALK status or number of prior therapies [156]. The estimated OS rate for the whole series and the subset of patients not achieving CR was 79% and 25%, respectively [156]. Moreover, the PFS rate among CR patients was substantially higher than in total enrolled patients (57% versus 39%, respectively). Similar results with an ORR of 62.5% (45% CR and 17.5% PR) were reported by an Italian observational, multicenter, retrospective study in 40 patients with systemic ALCL [157].
Activity of BV monotherapy in other types of R/R PTCL is inferior as compared to ALCL, with ORR being more frequently observed in 54% of AITL cases [140, 158]. The median duration of response was limited. BV has been also used in combination with bendamustine in R/R PTCL patients (n = 82) [159]. The best ORR was 68%, with 49% CR. Median duration of response was 15.4 months for patients in CR. The combination BV + ICE (ifosfamide, carboplatin, etoposide) in R/R PTCL was associated with an ORR of 66.7%, with all patients achieving a CR [160]. The activity of BV plus nivolumab in R/R PTCL and cutaneous T‐cell lymphoma has been more modest [161].

10
Other CD30 Expressing Lymphomas
CD30 is variably expressed in DLBCL and mediastinal lymphomas, and the activity and safety of BV have been documented in R/R DLBCL, both as monotherapy [162] with 44% ORR (including 17% CR) or in combination with lenalidomide, with 57% ORR and 35% CR [163]. The combination of BV + nivolumab was studied in R/R primary mediastinal B‐cell lymphomas. After a median follow‐up of 11.1 months, the ORR was 73%, with a CR rate of 37% [164]. Due to the high success of anti‐CD19 chimeric antigen receptor (CAR) T cells in R/R DLBCL and primary mediastinal lymphomas, there has been a decreased interest in BV in this setting. However, this agent may still have a role as bridge therapy to CAR‐T cells.

11
Anti‐CD30 CAR‐T Cells
11.1
Treatment of R/R cHL and Other CD30+ Lymphomas
CAR‐T cells directed against CD19 have revolutionized the treatment of R/R B‐cell acute lymphoblastic leukemia and B‐cell lymphomas [165, 166]. Results of CD30‐directed CAR‐T cells in R/R cHL have been more modest [167]. In the first published phase I trial [168], 18 heavily pretreated patients (17 with R/R cHL and 1 with ALCL) received anti‐CD30 CAR‐T cells. The ORR was 39% (all partial responders) and the median PFS was 6 months. All patients experienced grade 1–2 cytokine release syndrome (CRS) but not immune effector cell associated neurotoxicity syndrome (ICANS) or treatment‐related deaths. Other studies showed similar unsatisfactory responses [169, 170], associated with high toxicity [170].
In a phase I/II trial [171], 41 patients with heavily pre‐treated cHL received anti‐CD30 CAR‐T cells. The ORR of the whole cohort was 62% with a 1‐year PFS of 36%. Among the 32 patients who received a fludarabine‐based lymphodepletion, ORR was 72%, with 19 patients (59%) achieving CR; the 1‐year PFS and OS of this subgroup were 61% and 94%, respectively. CRS was of low grade, and no neurologic toxicity was observed. These favorable results were confirmed in 15 refractory cHL patients with a CR rate of about 60% [172, 173]. However, a high CR rate in cHL is usually associated, at a follow‐up of 6 years, with a low PFS of only 19% and a duration of response of 25% [174].
ASCT was used in tandem with anti‐CD30 CAR‐T cell infusion to treat R/R CD30+ lymphomas [175]. This approach is well‐tolerated and highly effective in R/R cHL and ALCL, even in PET‐positive or chemorefractory patients who are expected to have inferior outcomes after ASCT.
Third‐generation anti‐CD30 CAR‐T cells were evaluated in combination with lymphodepletion in R/R CD30+ lymphoma [176]. The median PFS for the 9 patients was 13 months, with three long‐term CRs over 2 years response. This may be explained by the two costimulatory domains CD28 and 4‐1BB [176] that can both facilitate CAR‐T cell proliferation (due to the CD28 costimulatory domain) and prolong CAR‐T cell persistence in vivo (due to the 4‐1BB endodomain) [177, 178]. In another study [179] CD30.CAR‐Ts co‐expressing the cognate receptor for CCL17, CCR4 (CCR4.CD30.CAR‐Ts) with improved tumor homing and anti‐lymphoma activity compared with CD30.CAR‐Ts not expressing CCR4 were used to treat 8 R/R cHL patients, with 6 (75%) achieving a CR and 2 (25%) a PR. Five patients are in remission to date, with one patient still in CR at 2.5 years post treatment.
Anti‐CD30 CAR T cells have also been used as consolidation after ASCT in CD30‐positive lymphoma patients at high risk of relapse [180]. In particular, 18 patients (11 cHL, six T‐cell lymphoma, one gray zone lymphoma) were infused with anti‐CD30 CAR‐T cells at a median of 22 days (range 16–44) after ASCT. One patient had grade 1 CRS. The most common grade 3–4 adverse events were lymphopenia and leukopenia (11% of cases). At a median follow‐up of 48.2 months post‐infusion, the median PFS for all treated patients (n = 18) was 32.3 months and the median PFS for treated patients with cHL (n = 11) has not been reached [180]. The median OS for all treated patients has not been reached.
Despite the high response rates to anti‐CD30 CAR‐T cells, when preceded by lymphodepleting chemotherapy, disease progression is common, with treatment failures being mainly correlated to higher metabolic tumor volume (> 60 mL) by PET before CAR‐T cell infusion [181]. Conversely, bridging therapy, anti‐CD30 CAR‐T cell expansion/persistence, and the percentage of CD3+ lymphocytes over the first 6 weeks of therapy did not impact PFS.
A recent meta‐analysis on the safety and efficacy of anti‐CD30 CAR‐T cell therapy in R/R cHL was performed on a total of 151 participants [182]. The ORR and CR were 57% and 34%, respectively. A partial response was observed in 32% of cases. With the median follow‐up range from 9.5 to 71.5 months, the 1‐year PFS was 39%, and the 1‐year OS was 89%. Leukopenia and CRS were the most common adverse events, but they were tolerable and resolved with treatment. Clinical trials with anti‐CD30 CAR‐T cells in cHL are summarized in Table 2.

11.2
Mechanism of Resistance to CD30‐Directed CAR‐T Cells and Treatment of Relapses After CAR‐T Cells
Mechanisms of resistance to anti‐CD30 CAR‐T cells have been investigated. Decreased CD30 expression after BV and anti‐CD30 CAR‐T cells was observed in a patient with cHL [184]. Similar findings were reported by Marques‐Piubelli [185] but these were not confirmed [186]. Conversely, other investigators found that CD30 expression is usually retained in relapsing tumors after anti‐CD30 CAR‐T cell therapy. These findings suggest that the recurrence may be more likely due to the low affinity binding scFv to CD30 and/or to the highly immunosuppressive tumor microenvironment in cHL. Thus, strategies have been adopted to overcome these problems. Artificial intelligence and surface plasmon resonance were used to select scFv fragments of mAbs with high affinity for CD30 [187]. Dual anti‐CD30/PDL1‐CCR CAR‐T cells can be adopted as a strategy. In this way, PDL1.CCR is expected to block PDL1/PD1 interaction of CAR‐T cells and macrophages with PDL1+ tumor cells, mimicking an immune checkpoint effect and reducing CAR‐T cell exhaustion [188]. At the same time, stimulation of PDL1‐4.1 BB costimulatory receptor would enhance CAR‐T cell proliferation and persistence [188].
In patients relapsing after CAR‐T cell treatment, PD‐1 inhibitors may re‐induce CR, even in patients previously exposed to or progressed under these drugs [176]. In addition, anti‐CD30 CAR‐T cells can be combined with anti‐PD1 antibody therapy [167, 183, 189]. In particular, the anti‐PD‐1 antibody enhanced the effect of CD30‐directed CAR‐T therapy in R/R CD30+ lymphoma patients, with minimal toxicities [189]. The ORR was 91.7% (11/12), with 6 patients achieving CR (50%), and CRS was of low grade. With a median follow‐up of 21.5 months (range: 3−50 months), the PFS and the OS were 45% and 70%, respectively.

12
Conclusions
The CD30 molecule is an excellent diagnostic marker for cHL and ALCL. Targeting CD30 is also valuable for lymphoma treatment. The high CR rates observed when incorporating BV in the frontline treatment of early‐stage cHL suggest that this strategy may potentially lead to omitting radiotherapy in this setting. Patients with advanced stage cHL who are not eligible for anthracycline‐based chemotherapy may benefit from BV‐nivolumab. The activity of BV as frontline therapy in some PTCL categories needs to be further investigated, especially given the variability of CD30 expression across subtypes, with the highest levels being found in ALCL and cHL. The optimal CD30 expression cutoff for patient selection remains an unsolved issue [140, 190]. Responses to BV have been observed with low or even absent expression of CD30, as in various subtypes of T‐cell lymphomas and mycosis fungoides [150, 191, 192]. Such responses could be due to the presence of low levels of surface CD30, below the threshold of detection by immunohistochemistry. A more appealing hypothesis is the bystander effect of BV, where MMAE crosses the cell membrane of the rare killed CD30+ tumor cells and is released into the surrounding extracellular matrix, exerting its cytotoxic activity on adjacent CD30‐negative tumor cells [193]. The activity of anti‐CD30 CAR‐T cells has been so far modest, and attempts should be made to modify them structurally, with the aim to improve the anti‐tumor activity and also to antagonize the immunosuppressive microenvironment of cHL.

Ethics Statement
The authors have nothing to report.

Conflicts of Interest
The authors declare no conflicts of interest.