Induction of Sustained Remissions Associated with Immune Activation by Idelalisib in Patients with Follicular Lymphoma

Abstract

Phosphatidylinositol-3-kinase (PI3K) inhibition has emerged as a therapeutic option against indolent lymphoma, including relapsed follicular lymphoma (FL). While inhibition of active signaling in the lymphoma cell represents the primary mode of action, PI3K inhibition also exerts immunomodulatory effects. Here we have analyzed 17 consecutive advanced treatment line FL patients treated with the delta-selective PI3K inhibitor idelalisib in a retrospective single-center observational study, with a specific focus on response and immune effects. Eleven patients achieved complete remission (CR) or partial remission (PR) with median response duration of 22 (11–88) months following a median idelalisib exposure of 15 (4–88) months. Disease response persisted in three patients for a median of 37 (21–63) months following cessation of idelalisib without another therapy being initiated. Autoimmune side effects occurred in eight of the eleven patients who responded, compared to none in six patients whose disease did not respond. In conclusion, a time-limited exposure to idelalisib may induce sustained remissions in a portion of patients with recurrent and/or refractory (r/r) FL, suggesting immunomodulatory effects of PI3K inhibition to be involved in the control of the disease.

1. Introduction

While the clinical course of FL is generally indolent, remissions induced by chemotherapy are mostly transient and the long-term outcomes for patients with chemotherapy-refractory disease remain poor. The PI3K signaling pathway is involved in multiple functions of both B- and T-lymphocytes and its aberrant activation contributes to the propagation and expansion of the transformed cells in several lymphoma entities [1,2,3]. Inhibition of the pathway perturbs PI3K-dependent signaling in the tumor cells, showing significant anti-lymphoma cytostatic activity in vitro and in vivo [4], and, based on its clinical activity, the highly selective PI3K delta (δ) inhibitor idelalisib [5] has been approved to treat patients with relapsed FL [6,7].

Autoimmune side effects of PI3Kδ inhibition are common, including hematotoxicity, fever, nausea, diarrhea, fatigue and elevated liver enzymes. Experimental evidence suggests that regulatory T-cells (Tregs) contribute to these immune-related toxicities. Mice with genetic inactivation of p110δ develop autoimmune colitis [8] associated with a reduction in Tregs [9], which has also been demonstrated in the peripheral blood of patients treated with idelalisib [10]. Idelalisib-associated diarrhea was associated with intraepithelial lymphocytosis [11]. In patients with idelalisib related hepatitis, liver biopsies showed infiltrates of activated T-cells. Idelalisib-associated hepatitis typically takes several weeks to develop following first exposure, but rapidly recurs after restarting the therapy, a finding compatible with an adaptive immune response [10]. Most adverse effects are reversible by dose reduction, immunosuppressive drugs such as corticosteroids, or temporary interruption of idelalisib [12].

Starting idelalisib therapy induces a transient surge in blood lymphocyte counts, due to disrupted lymphocyte trafficking [13]. In-depth studies showed that PI3K inhibition causes a reduction in Tregs (CD4 positive) most likely caused by decreased cytokine gene expression and an increase in cluster of differentiation 8 (CD8) positive T cells [14]. The reduction in Tregs in the peripheral blood was seen particularly in patients experiencing adverse side effects [10,15,16]. In-vivo mice models showed that PI3Kδ inhibition induced an adaptive anti-tumor immune response through inhibition of an immunosuppressive Treg population, leading to a ‘rebalancing’ of the adaptive immune system from CD4+ T-helper-cells and CD8+ T-cells towards CD8⁺ T-cells [17,18,19]. The immunologic effects of PI3Kδ inhibition may also apply to cancer, where the association of immunologic side effects and increased overall survival (OS) of patients has already been described for solid tumors and chronic lymphocytic leukemia [20,21].

Here we report our experience with response and response duration in 17 consecutive patients receiving idelalisib salvage therapy against progressive FL, suggesting an association of sustained response and immunological activity in a subset of this cohort.

2. Results

2.1. Study Cohort

The index cohort of 17 patients with FL disease comprised 9 females and 8 males with a median age of 64 (42–78) years at the start of therapy with idelalisib. All patients had received a median of 3.3 prior chemotherapy (Ctx) lines, including chemotherapy related to prior autologous and allogeneic stem cell transplantation (ASCT).

Idelalisib treatment induced CR or PR in 11 cases while the disease progressed in 6 patients following 100 days of treatment with idelalisib. Analogous to previous studies the main adverse side effects were gastrointestinal (GI) toxicity and haematoxicity. Further side effects were elevation of liver enzymes, alopecia, muscle cramps and skin rashes. During the study 10 patients passed away, 7 caused by progress of their lymphoma. By the end of our observation period idelalisib had already been stopped in the 7 remaining patients due to either progress or side effects while three patients remained in remission.

The median age at the time of commencement of idelalisib was 55 (42–78) years in female and 64 (45–72) years in male patients. The female patients reached an overall survival (OS) after start of idelalisib of 24 (10–96) months, as compared to a significantly shorter OS of 13.5 (3–56) months in male patients. The female cohort showed a progression free survival (PFS) of 13 (2–88) months while the male cohort reached 8.5 (2–56) months.

2.2. Response to Idelalisib

Patients with response to idelalisib showed a longer duration of response duration following prior therapies (24 compared to 11 months) and had received slightly more therapy lines previously (3.5 to 3), including high dose chemotherapy and ASCT (

Table 1. Patient characteristics according to response. Outcome of idelalisib treatment according to response at day 100. OS: overall survival from start of idelalisib, PFS: progression free survival from start of idelalisib; CTX: chemotherapy; HD CTX/ASCT: high-dose chemotherapy with autologous stem cell transplantation; PD progressive disease; AE: adverse events; time intervals given in months.

	Responder	Non-Responder	All
N	11	6	17
Median age at ED (range)	46 (33–60)	67 (56–74)	56 (33–74)
Median age at start idelalisib (range)	52 (42–72)	68.5 (64–78)	64 (42–78)
Gender (f/m)	7/4	2/4	9/8
FL grading 1,2	8	3	10
FL grading 3A	3	3	6
Stage I, II (initial)	5	2	7
Stage III, IV (initial)	6	4	10
Prior lines CTX median (range)	3.5 (2–5)	3 (1–6)	3.35 (1–6)
Prior bendamustin	5	4	9
Prior HD CTX/ASCT	5	1	6
Prior response duration, median	24	11	19
PD pre idelalisib	9	6	15
Elevated LDH pre idelalisib	5	3	8
Autoimmune AE	8/11	0/6	8/17
Stop idelalisib	11/11	6/6	17/17
Treatment duration, w pausing; median (range)	15 (4–88)	2 (2–7)	6 (2–88)
PFS; median, (range)	22 (7–88)	3 (2–7)	12 (2–88)
OS; median, (range)	24 (10–96)	10.5 (3–15)	15 (3–96)

). There was no difference between responders and non-responders with regard to the initial lymphoma stage at first diagnosis or their ECOG performance status when their diagnosis was first made. As expected, responders showed a significant longer OS [24 (10–96) to 10.5 (3–15) months, p-value: 0.0005)] and PFS [22 (1–88) to 3 (2–7) months, p-value: <0.0001] after start of idelalisib (Figure 1a,b). 72.7% of the responders showed autoimmune side effects, mostly diarrhea, while such side effects did not occur in non-responders. 6 patients had a history of prior autologous and/or allogeneic SCT. Considering responders and non-responders separately by gender we observed 7 responders in the female (77.7% of observed female patients) and 4 in the male cohort (50% of observed male patients). The median age of the responders was younger compared to the non-responders both at the time of first diagnosis and at the introduction of therapy with idelalisib. The median age of female responders was 30 years lower than the median age of female non-responders [46 (34–66) to 74 (74) years] at initial diagnosis and 26.5 years lower at the initiation of idelalisib therapy [50 (42–71) to 76.5 (71–78) years]. The male cohort showed a median age gap of 9.5 years between responders and non-responders [51 (33–66) to 60.5 (56–71) years] at the time of initial diagnosis and a difference of 8.5 years when idelalisib was commenced [58 (45–72) to 64.5 (64–72) years]. The female responders showed an OS of 124 (85–180) months considered from the initial diagnosis and an OS after start of idelalisib of 25 (10–96) months following initiation of idelalisib therapy. In comparison, we observed a shorter OS in the male responders with an OS after initial diagnosis of 73.5 (28–96) months and an OS of 21 (18–56) months following the start of treatment with idelalisib. The female responders showed a median PFS of 22 (7–88) months and the male cohort one of 19.5 (12–56) months. 85.7% of female responders experienced autoimmune side effects while only 50% of male responders suffered such side effects, mainly GI toxicity.

2.3. Role of Autoimmune Side Effects

Considering treatment responses with and without autoimmune side effects we detected a median OS after initial diagnosis of 110 (90–180) months in responders with autoimmune side effects and 57 (28–85) months in responders without such side effects, i.e., a difference of 53 months. Median OS after the start of idelalisib was 29.5 (10–96) months in patients with autoimmune side effects and 12 (3–25) months in the patients without (Figure 1c, p-value: 0.0149). Likewise we detected a longer PFS after start of idelalisib in patients with autoimmune side effects 17.5 (7–88) months in comparison to patients without autoimmune side effects 5 (2–24) months (Figure 1d, p-value: 0.1057). Response following temporary interruption of idelalisib persisted in patients with auto-immune side effects with a median PFS of 3.5 (0–63) months, compared to 0 (0–21) months in responders without autoimmune side effects, suggesting an association of auto-immune effects with response and response duration (Figure 2).

This study was retrospective in nature, thus no specific immune monitoring for idelalisib-mediated immunological effects was planned. However, routine monitoring of CD4/CD8 counts and routine imaging allowed the observations shown in Figure 3 and Figure 4. We detected strong expansion of CD8 cells in patient 16 (Figure 4), a 47-year-old female. In her course of treatment idelalisib had to be paused twice due to gastrointestinal side effects. Restarting idelalisib repeatedly elicited expansion of CD3+/CD8+ peripheral lymphocytes (Figure 3).

Equally, in patient 17, idelalisib treatment and response were associated with intestinal side effects (grade 1–3). Idelalisib was applied as a fifth line therapy against FL and PR and CR were achieved at month 3 and 9 respectively. Colonoscopy revealed colitis which responded promptly to corticosteroids. In this female 41-years old patient a thymus rebound was detected 58 months following the initiation of idelalisib therapy (Figure 4), the hyperplastic thymus was PET/CT negative and remained without progression through month 85.

2.4. Long-Term Response

Three patients experienced a continuing response under therapy with idelalisib, all with autoimmune associated GI side effects. Their median age at initial diagnosis was 45 years (41–60). With a median response of 85 (56–88) months we detected a significant difference to the median response of 22 (11–88) months in the entire responder cohort.

Likewise, three patients displayed a long-lasting disease response for a median of 37 months (21–63 months) even after cessation of idelalisib due to side effects. Two of them experienced GI-toxicity, one developed idelalisib associated exanthema and leukocytopenia. Two of them are part of the long-term responder group indicated above.

With the exception of one fatal event due to influenza infection we have not observed treatment-related deaths due to an infectious complications, adhering to standard prophylactic antibiotic recommendations and early intervention in the case of infection.

3. Discussion

PI3K inhibition has evolved as an effective treatment modality against r/r indolent lymphoma, leading to approvals for idelalisib and copanlisib in r/r follicular lymphoma [22,23]. In our retrospective single center cohort with special attention to long-term outcome idelalisib induced complete or partial remissions in 59% of patients, with a duration of response of 22 months (Table 1). It is noteworthy that three patients were still in remission at the latest follow-up at 96 months.

Intraindividual disease control induced by idelalisib persisted longer than the median of response durations induced by prior lines of chemotherapy (Table 1). This contrasts the generally accepted concept of conditional survival in general oncology, describing the shortening of response durations with increasing treatment lines of conventional cytostatic therapy [24]. This may suggest a contribution of indirect effects to the drug`s efficacy in some cases.

Preclinical evidence clearly established the disruption of B-cell receptor (BCR) signaling at the level of PI3K activation as the predominant mode of action of idealisib against B-cell lymphoma cells [6,25]. Systemic PI3K inhibition, however, also affects T-cells and in particular regulatory T-cells (Treg) [17,26,27]. Both in mouse models and in human ex vivo studies PI3K inhibition blocked inhibitory Treg activity, allowing an increased anti-tumor disease activity of effector T-cells [17,27]. Treg inhibition also explains the frequent autoimmune activity observed in clinical PI3K inhibitor trials, including idelalisib [28]. Importantly, recent studies of the T-cell repertoire in chronic lymphatic leucemia (CLL) patients revealed the occurrence of dominant T-cell clones associated with idelalisib treatment and suggested an association of T-cell repertoire shifts with anti-lymphoma effects [29].

Our findings suggest a connection between the response to treatment and immune-related side effects. Such effects were reported in 72.7% of our responders, but in none of the non-responders. Furthermore, responders who experienced auto immune side effects experienced a longer OS following the initiation of idelalisib therapy as compared to responders without autoimmune effects. This shows that immune toxicity may be a surrogate marker of efficacy in this drug class and an early predictor of durable response.

The thymic hyperplasia which we found in one patient with long-term response and autoimmune GI side effects may also hint towards immunomodulatory effects of PI3K inhibition. Thymic hyperplasia is a common phenomenon after Ctx for lymphoma in pediatric patients [30] but rarely occurs in older patients [31]. This is thought to be due to the fact that in adults over the age of 40 the thymus consists of fatty tissue with only low counts of immune cells [32]. Usually rebound can be detected within two years following the begin of chemotherapy [33]. Thymic hyperplasia can then be divided into two subtypes, true hyperplasia and lymphoid hyperplasia, which may be difficult to distinguish on imaging [33,34]. In true hyperplasia, the thymus most commonly appears oval, with a smooth contour, homogeneous and lack of contrast medium enhancement [33]. These features are all present in our case. However, a study on the imaging features of pathologically proven true thymic hyperplasia revealed that such qualitative features did not significantly differ between the subtypes [34]. It concluded, however, that CT attenuation of the thymus was higher in lymphoid than in true hyperplasia. The suggested cutoff value was 41.2 HU, a value not reached in any of the CT studies in our case. In addition, the same study found that SUV_max measurements in ¹⁸F-FDG PET/CT studies were below 3.8 in true hyperplasia, again concordant with the reading of 1.5 in our patient.

Although gained retrospectively from a limited number of patients our observations in r/r FL patients support the notion that idelalisib mediated T-cell activation participates in the drug’s anti-lymphoma effect (Figure 2). Admittedly, the precise mechanisms as well as (bio)markers for a prospective identification or monitoring of those patients benefitting remain to be elucidated. Further investigation appears warranted to identify those patients who may benefit most from PI3K inhibitor immune effects.

Therapeutic intervention with PI3K inhibition certainly will have to balance beneficial anti-tumor activity against side effects through auto-immune phenomena or opportunistic infections triggered by PI3K immunomodulatory effects [22]. In a wider perspective and with targeted immunological treatment options emerging [35,36], PI3K mediated Treg inhibition may evolve as a valuable tool in combined immunotherapy approaches. Most promising in perspective, PI3K delta inhibitor therapy appears a candidate to be combined with CAR-T or bi-specific antibody treatments, aiming at enhanced activity by downregulation of suppressive TREGs.

4. Materials and Methods

Data for this retrospective single-center observational study were extracted from medical records following our institution’s ethics review board approval. All patients who had received idelalisib according to EMA approval against advanced progressive FL between February 2015 and March 2023 were included. Disease monitoring adhered to good clinical practice applying the Lugano criteria [37], end point for response evaluation was day 100 following the initiation of treatment. Primarily, overall survival (OS) and progression-free survival (PFS) were measured in our patient cohort. OS and PFS were compared between subgroups using Kaplan-Meier curve. Statistical significance was tested with log-rank test. Adverse events as recorded in the patients’ charts were documented with special regard to immune side effects and their appearance in the subgroups.

The majority of patients started with idelalisib at a standard dose of 150 mg twice daily, with dose reductions in the course of treatment according to drug tolerability.

5. Conclusions

Our observations support a judicious use of idelalisib in patients with r/r FL: once the drug induces remission, a chance for lasting benefit emerges, even upon drug cessation. Further studies may show how to integrate PI3K inhibition optimally in lymphoma immunotherapy.