Successful PD-1 inhibitor treatment in a patient with refractory transformed follicular lymphoma who failed to respond to CAR-T cell therapy: a case report and literature review

ABSTRACT

Follicular lymphoma (FL) accounts for approximately 35% of all non-Hodgkin lymphomas and can progress to diffuse large B cell lymphoma (DLBCL) at a rate of 2% per year. Here, we present a 56-year-old female patient who was diagnosed with grade 3a FL. Further pathological investigation revealed that the lymphoma had transformed into DLBCL following six courses of R-CHOP regimen, and further disease progression was observed after two courses of R2-GemOx. We ultimately failed to collect hematopoietic stem cells after two courses of R2-ICE. CD-22 CAR-T cell therapy salvaged the patient; however, a new soft tissue mass of 4.8 × 4.1 cm rapidly emerged in the patient’s right lung. Following the observation of strong tissue staining of PD-L1 (90%), the patient was administered PD-1 inhibitor and 26 Gy of radiotherapy and has maintained progression-free survival at more than 15 months of follow-up. Transformed FL with strong PD-L1 expression showed a poor response to standard immunochemotherapy. Our findings support the potential benefit of PD-1 inhibitor and combination therapies in this type of transformed FL.

KEYWORDS:

• Transformed follicular lymphoma
• PD-1 inhibitor
• CAR-T
• refractory
• diffuse large B cell lymphoma

Introduction

Follicular lymphoma (FL) accounts for ~20-25% of all new non-Hodgkin lymphoma (NHL) diagnoses in western countries¹ and is characterized by the occurrence of t(14;18) translocation, resulting in the over-expression of anti-apoptotic BCL2 protein.² Early recurrence after first-line chemotherapy occurs in about 20% of FL patients.³ FLs in general do not express PD-L1, which may explain the poor efficacy of PD-1 inhibitors observed in this patient population. In a phase 2 study of nivolumab for relapsed/refractory FL followed up for 12 months, an overall response rate (ORR) of 4% (4/92) and median progression-free survival (PFS) of 2.2 months was reported.⁴ Anti-PD-1 monoclonal antibody therapy may therefore be associated with very limited activity in patients with relapsed or refractory (R/R) FL. Furthermore, approximately 28% of FL will transform into diffuse large B cell lymphoma (DLBCL) within 10 years,⁵ with a risk of 2% per year.² The curative effect of advanced-stage follicular lymphoma undergoing histologic transformation is not satisfactory. The present case report describes a meaningful alternative treatment strategy for patients with refractory tFL treated with an anti-PD-1 monoclonal antibody and combination therapies.

Case presentation

A 56-year-old woman presented with aggravated abdominal pain and night sweats with a duration of 3 days in September 2018. Laboratory test results revealed elevated $\beta$2-MG of 4 $\mu$g/mL and extremely elevated serum lactate dehydrogenase at a level of 684 U/L. Epstein Barr virus (EBV)-DNA in whole blood was 8371 copies/mL and lymphocyte subsets showed that B-cell counts were extremely low at 5/μL. Computed tomography (CT) scanning of the abdomen revealed multiple enlarged lymph nodes. Subsequent positron emission tomography (PET)/CT scanning revealed the presence of multiple enlarged abdominopelvic cavities and retroperitoneal lymph nodes, the largest of which was 40 × 57 mm in the peripancreatic region with a high standardized uptake value of 41.9 (Figure 1a). PET/CT scanning has also indicated hepatosplenomegaly, with fluorodeoxyglucose (FDG) uptake significantly increased in multiple bone lesions. Histologic analysis of a core needle aspiration biopsy of the lymph node with the highest FDG uptake resulted in the following: CD20+ (Figure 2a), CD10-, BCL6+, Ki67 (70%+), C-MYC (40%+), MUM1+, CD21+ (Figure 2b), BCL2+ (Figure 2c), CD38-, CD5-, CD3-, EBER-, and CKpan-. Fluorescence in situ hybridization indicated a negative rearrangement of BCL2 and MYC. Bone marrow smear was normal. In consideration of the patient’s symptoms and immunohistochemistry findings, a diagnosis of FL grade 3a (stage IV, group B) with a high risk of FLIPI 3 points was made.

Figure 1. Whole-body PET/CT scan prior to treatment initiation (a). PET/CT imaging showed several enlarged or new lesions after six cycles of R-CHOP (b) and following CAR-T cell therapy (c)

Figure 2. Immunohistochemical (IHC) staining of three tissue biopsy samples. IHC staining of CD20 (a), CD21 (b), and BCL2 (c) in the first tissue biopsy samples. Positive IHC staining of CD20 (d) and MUM-1 (e) and negative staining of CD10 (f) in the second tissue biopsy samples. Positive staining of CD20 (g), CD22 (h), and PD-L1 (i) in the third tissue biopsy samples. (original magnification, 200×)

PET/CT after six courses of R-CHOP showed a clear response with metabolic normalization of multiple bone lesions; however, several new lesions appeared. A new left axillary lymph node was 14 × 15 mm with a high FDG uptake value of 37.8 (Figure 1b). Immunohistochemistry of the second pathological tissue in the left axillary lymph node revealed that FL had transformed into DLBCL (Figure 2). Fluorescence in situ hybridization indicated a negative rearrangement of BCL2 and MYC and a positive rearrangement of BCL6. The patient was subsequently diagnosed with DLBCL of non-germinal center B-cell type (stage III, group B), age-adjusted IPI score 2, high-intermediate risk. Disease progression was observed after two courses of R2-GemOx. Hematopoietic stem cell transplantation (HSCT) was considered as the patient had achieved partial remission (PR) after two courses of R2-ICE; however, this treatment strategy was not pursued due to failure of hematopoietic stem cell mobilization. After fludarabine and cyclophosphamide as lymphodepleting chemotherapy, the patient received approximately 5 × 10⁶/kg of anti-CD22–4-1BB–CD3ζ chimeric antigen receptor T (CAR-T) cells. We used flow cytometry to detect the CD22-CAR. CAR-T cells expression in peripheral blood CD3+ cells increased to 30% in the first 10 days after infusion, then decreased gradually. On day 30 after CAR-T cell infusion, whole-abdomen enhanced CT scanning showed the patient developed progressive disease (PD) according to the Lugano classification. PET/CT evaluation on day 43 after CAR-T cell infusion discovered a significant increase in the emerging lesions, especially the right pulmonary hilum lymph node where a lesion of 4.8 × 4.1 cm in size with maximum standard unit value (SUV_max) of 32.4 was observed (Figure 1c). And at this time, we measured that the expression of CAR-T cells was decreased significantly. Given the expression of BCL2, the patient was administered venetoclax, but no response was observed. Core needle lung puncture on day 55 after CAR-T cell infusion revealed that the expression of PD-L1 was extremely high (90%) (Figure 2i). No significant mutations were discovered in second-generation sequencing. Considering the PD-L1 expression level, the PD-1 inhibitor sintilimab (200 mg) was administered on day 76 after the infusion and CAR-T cells were almost undetected at that time. Fortunately, the patient achieved partial remission (PR) after the first cycle of PD-1 inhibitor. Ten days later, 27 times radiotherapy (2 Gy each time) was carried out but was subsequently stopped because of grade 4 hematologic toxicity when the patient completed 26 Gy. Encouragingly, the patient has achieved sustained complete response (CR) with only PD-1 inhibitor treatment with about 15 months after cessation of radiotherapy (Figure 3). PD-1 inhibitor and concurrent radiotherapy may mutually promote the efficacy in the treatment of this case, and the complete response might be considered as a result of combination therapies.

Figure 3. CT scan prior to initiation of PD-1 inhibitor treatment (29 Oct 2019) (a), one cycle after PD-1 inhibitor treatment (11 Nov 2019) (b), and two cycles after treatment initiation (9 Dec 2019) (c). The patient remains in CR by maintenance therapy with PD-1 inhibitor to date (d)

Discussion

FL can progress to aggressive lymphoma at an estimated rate of 2% per year, with the most common type of transformed lymphoma is DLBCL.⁶ The majority of tFL cases (80%) have been shown to derive from the germinal-center B-cell–like subtype, while a considerable proportion (approximately 16%) derive from the activated B-cell–like (ABC) subtype. Of note, ABC subtype cases typically originate from FLs without BCL2 translocation,⁷ which may in part contribute to the over-expression of ABC-like and late GCB cell phenotype signatures in t(14;18)-negative FL.⁸ Previous studies have demonstrated that a key difference between patients with and without BCL2 rearrangement is that in t(14;18)-negative FL, new immunoglobulin N-glycosylation sites within the variable region of surface immunoglobulin (sIg) are rarely acquired during the somatic mutation of FL, which plays an important role in activating B-cell receptor (BCR) signaling bypass antigen presentation but interact with mannose-binding lectins expressed by macrophages and dendritic cells.^9,10 Such a mechanism may result in activating BCR signaling by classical antigen presentation in t(14;18)-negative FL. Furthermore, anomalous signaling originating from tumor microenvironment or NFkB-associated genes, leading to the significant augmentation of gene mutations associated with immune response (IR) /inflammation or NFkB-signaling may also be an important mechanism driving pathogenesis of t(14;18)-negative FL.¹⁰ Moreover, FL without BCL2 rearrangement is characterized by high expression of Ki67 and MUM1 and significant downregulation of CD10, with the latter event known to be associated with the absence of BCL2 expression in advanced-stage FL (p = .016).¹¹ These subtle but significant differences, including significant reduction in CD10 expression, enrichment of NF-kB signaling, and down-regulation of distinct microRNA profiles (e.g. miR-16, miR-26a, miR-101, miR-29 c and miR138), have been reported in t(14;18)-negative FL and are indicative of a “late” germinal center B-cell molecular phenotype in t(14;18)-negative FL.^10–12 This finding is consistent with the observation that t(14;18)-negative FL typically transforms into ABC-DLBCL.⁷ Although t(14; 18)-negative FL has some specific molecular characteristics, in terms of initial therapeutic efficacy, no difference has been observed in survival duration or overall survival between FL patients with and without BCL2-break/t(14;18).¹¹ However, another study considered CD10-MUM1+ FL to be different from typical FL in molecular features and prognosis. CD10-MUM1+ FL were almost negative for BCL2 rearrangement, showed poor prognosis compared with CD10+ MUM1- FL, and 88% were accompanied by translocation or amplification of BCL6 gene.¹³

FL patients with histological transformation (HT) have poor prognosis and short survival.¹⁴ In the present case report, the patient with transformed DLBCL demonstrated an unsatisfactory response to various chemotherapy regimens. A previous study has reported that the CR rate of tFL patients after salvage chemotherapy was lower (50.3% vs 67.4%) and the disease progression rate was higher (28.2% vs 9.6%) compared with patients without transformation, and the median overall survival was 3.8 years.¹⁵ Other studies have reported a 3-year overall survival (OS) rate of 44% among patients with early-stage FL undergoing HT, while those treated with R-CHOP at transformation had a 5-year OS rate of 66%.^6,16 A poor curative effect was observed in our patient following multiple immunochemotherapy, and PD-1 inhibitor was subsequently initiated on the basis of high PD-L1 expression in the patient’s first tumor tissue biopsy.

Several studies have shown that PD-L1 is rarely expressed in FL tumor cells.^17,18 In terms of therapeutic effect, in a phase 1 study of 10 R/R FL patients who were treated with nivolumab and had different levels of PD-L1 expression in nonmalignant immune cells within the tumor microenvironment, 1 CR (10%) and 3 partial responses (30%) were observed.¹⁹ In a phase 2 trial of 29 relapsed FL patients, 19 (66%) obtained an objective response with a CR in 15/29 (52%) patients and a PR in 4/29 (14%) patients following treatment with pidilizumab combined with rituximab.²⁰ The efficacy of PD-1 inhibitor treatment in DLBCL remains unsatisfactory. In a phase 2 study, 121 R/R DLBCL patients who were unsuitable for or had failed with autologous hematopoietic cell transplantation (auto-HCT) received nivolumab 3 mg/kg every 2 weeks. ORR was 10% in the auto-HCT-failed cohort and 3% in the auto-HCT-ineligible cohort, while median PFS was 1.9 and 1.4 months and OS was 12.2 and 5.8 months, respectively.²¹ A few clinical trials of PD-1/PD-L1 blockade as monotherapy or combination therapy in FL, tFL, or DLBCL patients are currently underway (Table 1). We observed strong expression of PD-L1 in the patient’s first tumor tissue, and inferred from this finding that immune evasion might play a decisive role in tFL disease progression. We propose several factors which may have contributed to the high expression of PD-L1 in this case. First, tumor cells, antigen-presenting cells (APC), and myeloid suppressor cells (MDSC) in the tumor microenvironment can upregulate PD-L1 expression when stimulated by inflammatory cytokines during antitumor immune responses, leading to adaptive immune resistance.^22,23 Furthermore, EBV-related FL has a higher rate of transformation to DLBCL; however, the relationship between EBV and FL remains poorly understood.²⁴ EBV infection may also play an important role in the development and/or progression of some FLs, and several studies have demonstrated elevated PD-L1 expression in EBV-infected lymphomas such as CHL and DLBCL.^25,26 In this case, the EBV-DNA was significantly increased when the patient was diagnosed with follicular lymphoma for the first time. And at that time, lymphocyte subsets showed that B-cell counts were extremely low at 5/μL. Therefore, we doubted that EBV may have a disfunction effect on the immune system. However, the correlation between EBV and the type of lymphoma in this patient remains unclear.

Table 1. Ongoing clinical trials of PD-1/PD-L1 blockade as monotherapy or combination therapy in FL, tFL, or DLBCL

Trial interventions	Malignancy	Status; estimated completion date	ClinicalTrials.gov NCT reference
Nivolumab (anti-PD-1 antibody)
Nivolumab	R/R DLBCLR/R FL. etc	Recruiting; August 2022	NCT04205409
Nivolumab	FL	Recruiting; June 2024	NCT03245021
Nivolumab, Chemotherapy	tFL DLBCL. etc	Recruiting; June 2022	NCT03704714
Nivolumab, Duvelisib	tFL R/R DLBCL. etc	Recruiting; December 2022	NCT03892044
Nivolumab, Copanlisib	FL DLBCL. etc	Recruiting; January 2023	NCT03884998
Nivolumab, Poly ICLC	FL	Recruiting; January, 2029	NCT03121677
Pembrolizumab (anti-PD-1 antibody)
Pembrolizumab	FL B-NHL. etc	Recruiting; September 2024	NCT03498612
Pembrolizumab and Rituximab With or Without Lenalidomide	R FL R DLBCL	Recruiting; November 2021	NCT02446457
Pembrolizumab and Vorinostat	R/R FL R/R DLBCL. etc	Recruiting; December 2021	NCT03150329
Pembrolizumab, Rituximab, and Obinutuzumab	R/R FL R/R DLBCL	Recruiting; February 2023	NCT03401853
Pembrolizumab Alone or With Idelalisib or Ibrutinib	R/R FL. etc	Recruiting; January 2023	NCT02332980
Pembrolizumab Combined With SIIT	FL	Recruiting; January 2024	NCT02677155
Atezolizumab (anti-PD-L1 antibody)
Atezolizumab-Venetoclax-Obinutuzumab Combination	FL DLBCL. etc	Recruiting; November 2021	NCT03276468
Atezolizumab, Gemcitabine, Oxaliplatin, and Rituximab	tFL R/R DLBCL	Recruiting; June 2021	NCT03321643

PD-1, programmed cell-death protein 1; PD-L1, programmed cell-death 1 ligand 1; FL, follicular lymphoma; tFL, transformed follicular lymphoma; DLBCL, diffuse large B cell lymphoma; R/R, relapse and refractory; R, recurrence; B-NHL, B-cell non-Hodgkin’s lymphoma.

Immunosuppressive checkpoint inhibitors and CAR-T cell therapy are approved treatments for a variety of hematological malignancies. Several studies have reported the efficacy of CAR-T cell therapy in patients with FL and tFL. Among 14 patients with R/R FL and 14 with R/R DLBCL receiving anti-CD19 CAR-T cell therapy, CR occurred in 71% of patients in the FL cohort and in 43% in the DLBCL cohort. Moreover, 89% of responding FL patients and 86% of responding DLBCL patients had sustained remission at a median follow-up of 28.6 months.²⁷ A phase 1/2 clinical trial in 8 R/R FL and 13 tFL patients demonstrated CR of 88% with FL and 46% with tFL following the infusion of 2 × 10⁶ CD19-CAR-T cells/kg.²⁸ However, few reports to date have evaluated the clinical outcome of CAR-T cell therapy combined with anti-PD-1 antibody therapy in B-NHL. In a single-center study among 11 R/R B-NHL patients who received a single dose of 3 mg/kg nivolumab on the third day after receiving a median of 8 × 10⁶/kg CD19 CAR-T cells infusion, ORR of 81.8% was observed, with CR and PR of 45.5% and 36.3%, respectively.²⁹ More importantly, all reported toxicities were evaluated as manageable and temporary. In another case of refractory FL treated by CD19 CAR-T cell infusion combined with a reduced dose of nivolumab (1.5 mg/kg), complete remission for 16 months was reported at the time of publication.³⁰ After CAR-T cell therapy failure, we determined that CD22 expression remained positive, ruling out antigen loss as a contributing factor. Given the strong expression of PD-L1 in this unique tFL patient, we speculate that PD-L1 over-expression may lead to T cell dysfunction, in line with findings from a previous study.³¹ Furthermore, CAR-T cells may produce pro-inflammatory cytokines which can contribute to increased expression of PD-L1.³² Taken together with the observations from our present case study, these earlier findings support the efficacy of therapeutic strategies which block the inhibitory interaction of PD-1/PD-L1 to enhance anti-tumor CAR-T cells efficacy in this patient population. A phase 1/2 study of pembrolizumab in patients who failed to respond to or relapsed after anti-CD19 CAR-T cell therapy for R/R CD19+ lymphomas including FL is in progress (NCT02650999). Recently, modified CAR-T cells which secrete PD-1-blocking single-chain variable fragments (scFv) have been shown to improve the anti-tumor activity of CAR-T cells and bystander tumor-specific T cells with low toxicity, as secreted scFvs remained localized to the tumor.³³ This CAR-T modification appears to be more effective than combination therapy. However, further clinical studies are needed to evaluate the safety and efficacy of combination therapy or ‘armored’ CAR-T cell therapy in transformed FL.

In conclusion, we report successful PD-1 inhibitor treatment in a patient with refractory tFL who failed to respond to CAR-T cell therapy. Standard tFL therapies are characterized by aggressive clinical courses and poor outcomes which may be attributable to immune evasion. The findings of the present case report indicate that clinical benefit in this type of FL may depend on the synergistic effect of PD-1 inhibitors and combination therapies instead of traditional multi-protocol chemotherapy, and highlight the potential activity of monotherapy PD-1 inhibitors in FL with high expression of PD-L1. However, further studies are required to confirm whether PD-1 inhibitors alone and in combination therapy are more effective for the treatment of FL compared with existing treatments, and whether the level of PD-L1 expression affects the treatment response and survival rate.

Informed consent statement

Informed written consent was obtained from the patient for publication of this report and any accompanying images.

Disclosure statement

No potential conflicts of interest were disclosed.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website

Additional information

Funding

This study was funded by the National Natural Science Foundation international cooperation [81570184], the Science and Technology Project of Nantong City [MS22018008], the Science and Technology Project of Nantong City [MS12017003-2], China Postdoctoral Science Foundation [2019M660127], Jiangsu Province Postdoctoral Science Foundation [2019K062], Jiangsu Province Postdoctoral Foundation [2019Z146].