A phase II open label, randomized clinical trial of atezolizumab with or without human recombinant IL-7 (CYT107) in advanced urothelial cancer

Abstract

Purpose

Advanced urothelial cancer generally has high mortality despite modern anti-PD-1/L1 antibody-based combinations. Augmenting checkpoint inhibitor-mediated immune responses with lymphocyte growth factors may improve outcomes. We conducted a randomized phase II study (CITN-14) in 47 patients to explore whether human recombinant IL-7 (CYT107) could be safely combined with PD-L1 inhibition to enhance responses.

Patients and Methods

Patients with urothelial cancer following platinum chemotherapy were randomized to atezolizumab alone or with CYT107 weekly for four doses. The primary objective was clinical efficacy by objective response rate (ORR). Secondary objectives included safety, toxicity, and other clinical outcomes. Correlative endpoints included peripheral immunophenotyping and quantification of cytokines.

Results

CYT107 plus atezolizumab was well-tolerated, without dose-limiting toxicities (DLTs), and lower grade 3–4 treatment-related adverse events (TRAEs) compared to atezolizumab. The ORR was 26.3% for the combination versus 23.8% for atezolizumab alone (p = 0.428). The complete response (CR) rate was 10.5% for the combination versus 4.8% for monotherapy. Three patients on the combination had responses >21 months versus one with monotherapy. CD4⁺ and CD8⁺ T lymphocyte expansion occurred in patients with response to combination therapy, with the greatest effect in T memory stem cell (T_scm) cells. Responding patients had elevated baseline CCL4 and decreased VEGF-A and TNF.

Conclusions

Combining CYT107 with atezolizumab was safe and resulted in lymphocyte expansion, a doubling of the CR rate, and durable responses exceeding 2 years, however, the ORR was similar to atezolizumab alone. Increased and sustained doses of CYT107 coupled with patient selection strategies should be further investigated.

Statement of Translational Relevance

This is the first study to evaluate the IL-7 agonist CYT107 in combination with an immune checkpoint inhibitor. We tested the combination with atezolizumab versus atezolizumab alone in urothelial cancer. A higher frequency of complete and durable responses was observed, but the combination did not improve objective response rate. Correlative studies showed vigorous expansion of naïve, effector and memory subsets of both CD4⁺ and CD8⁺ T lymphocytes, especially T_scm cells. The expansion of PD-1 low T_scm cells was most pronounced in patients on the combination arm with response to therapy and suggests testing an increased dose level of CYT107. These data will be informative to future studies with T cell augmenting therapy, which should incorporate patient selection strategies to identify T cell inflamed tumors or include additional therapies targeting immune exclusion from the tumor microenvironment.

Introduction

Urothelial cancer remains a common and lethal malignancy despite the rapidly evolving therapeutic landscape¹. Multiomic profiling has revealed that a proportion of urothelial cancers possess a T cell-inflamed tumor microenvironment and a high prevalence of neoantigens^2–5. Thus, clinical treatment with PD-1/L1-targeted immune checkpoint therapy can lead to deep and durable responses in patients. Immune checkpoint inhibitor therapy has become incorporated into treatment algorithms since 2016, when the anti-PD-L1 antibody atezolizumab became the first to receive regulatory approval in urothelial cancer. Ultimately five other anti-PD-1/L1 antibodies were approved, but the objective response rates remained only 20–25% in this cancer, regardless of line of therapy^6–9.

The use of anti-PD-1/L1 immune checkpoint inhibition was further optimized through combination therapies or use in the maintenance setting after chemotherapy^10,11. Most recently, outcomes in the front-line setting were drastically improved when pembrolizumab was given with the antibody drug conjugate enfortumab vedotin¹². However, even with this highly active combination, the median survival remains shorter than three years. Thus, there remains a need for continued development of novel strategies to improve clinical responses to immunotherapy.

IL-7 is an endogenous cytokine with constitutive expression in lymphoid tissues, which binds the IL-7 receptor, a dimer of an alpha chain (CD127) and gamma chain (CD132)¹³. The IL-7 receptor is expressed on both naive and effector T lymphocytes. Upon IL-7 binding its receptor a cascade of intracellular signals is triggered, including STAT activation, which results in T cell maturation, survival, and activation^13–15. CYT107 is a glycosylated recombinant human IL-7 (rhIL7) developed for weekly dosing. In the context of cancer, it has been developed with the intent to augment the anti-tumor immune response by enhancing the function of effector T cells and their prevalence in the tumor microenvironment. Indeed, CYT107 has been shown to result in peripheral lymphocyte expansion and increase cytokine production in cancer patients^16,17.

Given the relatively low response rates to anti-PD-1/L1 monotherapy in advanced urothelial cancer, we hypothesized that augmentation of effector T cell function with CYT107 would lead to improved clinical outcomes, such as higher objective response rate or longer duration of response. In this context, we conducted a randomized, open label Phase II study to test the lowest active dose level of CYT107 with atezolizumab versus atezolizumab alone in advanced urothelial cancer in patients with progression after front-line standard of care platinum-based chemotherapy. This is the first study of CYT107 in combination with immune checkpoint blockade.

Patients and Methods

Study design and patient selection

CITN-14 (NCT03513952) was an open-label, multi-center, randomized Phase II trial comparing treatment with atezolizumab alone versus atezolizumab plus CYT107 (rhIL7) in advanced urothelial cancer. Eligible patients had histologically confirmed metastatic urothelial carcinoma, including a primary site in the renal pelvis, ureters, urinary bladder, or urethra. Mixed histology was allowed if the predominant histology was urothelial carcinoma. Patients were required to have received prior platinum-based chemotherapy with documented recurrent or progressive disease. Cisplatin or carboplatin-based chemotherapy were allowed. Key eligibility criteria included ECOG performance status ≤2, creatinine clearance ≥30 mL/min/1.73 m2, and measurable disease per RECIST v1.1. Patients with HIV were allowed on study if stable on therapy with a CD4 count above 250. Patients were excluded if they had prior anti-PD-1/L1 therapy or >2 systemic cytotoxic chemotherapy regimens, not counting neoadjuvant or adjuvant chemotherapy given ≥12 months prior to recurrence. Treatment with an investigational agent within 4 weeks of initiation of study therapy was excluded. Other exclusion criteria included treatment with systemic immunostimulatory or immunosuppressive agents, known central nervous system malignancy, or autoimmune disease. Written informed consent was obtained from all patients prior to any study related procedures. The study was conducted in accordance with the principles of the Declaration of Helsinki and approved by an institutional review board.

Study objectives

The primary objective was to determine the clinical efficacy of the investigational treatment combination as determined by objective response rate (ORR), defined by Complete Response (CR) or Partial Response (PR) as measured by RECIST v1.1. Secondary objectives included safety and toxicity of CYT107 added to atezolizumab, duration of response (DOR), and overall survival (OS). Exploratory objectives were to determine the immune correlates of the clinical activity of the investigational treatment combination, including the number of tumor-infiltrating lymphocytes and interferon gamma associated gene signatures in the tumor microenvironment.

Study procedures and assessments

The study included a safety run-in phase during which six patients received combination therapy with staggered enrollment. The first cycle was 28-days to enable a 1-week lead in for CTY107. During C1 patients received intramuscular 10 mcg/kg CYT107 IM on days 1, 8, 15, and 22 and atezolizumab on day 8. Cycle 2 and after were 21 days with 1,200 mg intravenous atezolizumab alone given on day 1. A cohort of 3 patients was enrolled during the safety run-in and assessed for dose-limiting toxicities (DLTs) during a 28-day window. If 0–1 DLTs were observed, another 3 patients would be enrolled. A DLT was defined as any adverse event occurring within the first 28 days that was considered at least possibly related to the protocol treatment (atezolizumab, CYT107, or both) that met protocol defined criteria.

Once the safety run-in was complete, patients were randomized 1:1 between combination CTY107 plus atezolizumab vs. a control arm with atezolizumab alone. Patients on the control arm were treated on 21-day cycles and received 1,200 mg intravenous atezolizumab on day 1. Radiologic tumor assessments by CT or MRI were performed within 28 days of C1D1 and every 9 +/− 1 week while on treatment.

Statistics

The patient safety run-in was intended to demonstrate the incidence of the DLTs being ≤ 16.7%. These patients were not included in the overall analysis since they were not randomized. After the safety run-in, patients were randomized 1:1 and included an interim futility analysis that was conducted when 50% of the planned sample size had their first disease assessment with survival information. The null hypothesis was an ORR of 14.8%, which is the expected activity of atezolizumab monotherapy. Cochran Mantel Haenszel test was conducted to compare ORR in the experimental arm (atezolizumab + CYT107) to control arm. A one-sided significance level of 0.10 was considered significant for the test.

Correlative Studies

Clinical cell counts:

CBC testing to determine absolute lymphocyte counts, and CD4 and CD8 T cell counts were performed on longitudinally collected, locally processed blood samples at each participating clinical site using Clinical Laboratory Improvement Amendments (CLIA)-certified assays.

Whole Blood Flow cytometry:

Whole blood was collected into Sodium Heparin vacutainers at specified time points for the collection of Peripheral Blood Mononuclear Cells (PBMCs). Samples were collected and shipped ambient to the CITN’s Central Immune Monitoring Laboratory (CIML) so that the samples were received and processed within 24 hours of blood draw. The change in absolute numbers, frequencies, and phenotypes of T cells and other PBMC subsets were evaluated by 21-color cytometry. Whole blood was stained (100µL, 24hrs after collection) in Trucount tubes (BD Biosciences Cat# 662415, RRID:AB_2870547) with a validated 21-color antibody panel staining mix containing 1x Brilliant Stain Buffer (BD 563794), FACS Buffer (2% FBS in DPBS (Gibco 14190144) and the following previously titrated antibodies: CD3 BUV805 (BD Biosciences Cat# 612895, RRID:AB_2870183), CD4 BUV661 (BD Biosciences Cat# 612962, RRID:AB_2870238), CD8 BUV496 (BD Biosciences Cat# 564804, RRID:AB_2744460), CD11c AF700 (BD Biosciences Cat# 561352, RRID:AB_10612006), CD14 BV510 (BioLegend Cat# 301831, RRID:AB_10897803), CD16 BUV737 (BD Biosciences Cat# 612786, RRID:AB_2833077), CD19 BUV563 (BD Biosciences Cat# 565697, RRID:AB_2744312), CD25 BB515 (BD Biosciences Cat# 565096, RRID:AB_2739065), CD28 B700 (BD Biosciences Cat# 745905, RRID:AB_2743332), CD45 BUV395 (BD Biosciences Cat# 563792, RRID:AB_2869519), CD45RA vioGreen (Miltenyi Biotec Cat# 130-113-369, RRID:AB_2726139), CD56 AF647 (BioLegend Cat# 362513, RRID:AB_2564086), CD123 BV650 (BioLegend Cat# 306019, RRID:AB_11218792), CD127 PE-Cy5 (BioLegend Cat# 351324, RRID:AB_10915554), CD152 BV786 (BD Biosciences Cat# 563931, RRID:AB_2738491), CD197 APC/Fire 750 (BioLegend Cat# 353246, RRID:AB_2750147), CD223 PE (BioLegend Cat# 369306, RRID:AB_2629592), CD279 PE-Cy7 (BD Biosciences Cat# 561272, RRID:AB_10611585), CD366 BV421 (BioLegend Cat# 345007, RRID:AB_10900073), HLADR BV605 (BioLegend Cat# 307640, RRID:AB_2561913), TIGIT PE-D564 (BioLegend Cat# 372716, RRID:AB_2632931). The samples were incubated at room temperature for 15 minutes in the dark. Subsequently, 900 µL of 1x FACS Lysing Buffer (BD Biosciences Cat# 349202, RRID:AB_2868862) was added to stained samples for 15 minutes at room temperature before freezing at −80°C. Thawed samples were collected in batches on a BD Symphony A1 and analyzed using FlowJo (RRID:SCR_008520) v10.8 software.

PBMC Immunophenotyping:

PBMCs were isolated from the above-described whole blood samples using Ficoll-Hypaque and cryopreserved in LN2. PBMCs were analyzed using 14-color flow cytometry. Samples were thawed in RPMI 1640 containing 10% Human Serum, 2mM L-glutamine, 100 Units Penicillin-Streptomycin, and 50U/mL Benzonase. 0.5—1×10⁶ thawed PBMCs were used for immunophenotyping. To evaluate cellular viability, the samples were first stained with Live/Dead Fixable Blue Dead Cell Stain Kit as per manufacturer’s instructions (Molecular Probes L23105). Subsequently, the cells were stained with a validated 14-color antibody panel staining mix containing 1x Brilliant Stain Buffer (BD 563794), FACS Buffer (2% FBS in DPBS (Gibco 14190144) and the following previously titrated antibodies: CD3 BUV805 (BD Biosciences Cat# 612895, RRID:AB_2870183), CD4 BUV661 (BD Biosciences Cat# 612962, RRID:AB_2870238), CD8 BUV496 (BD Biosciences Cat# 564804, RRID:AB_2744460), CD28 BB700 (BD Biosciences Cat# 745905, RRID:AB_2743332), CD45 BUV395 (BD Biosciences Cat# 563792, RRID:AB_2869519), CD45RA vioGreen (Miltenyi Biotec Cat# 130-113-369, RRID:AB_2726139), CD127 PE-Cy5 (BioLegend Cat# 351324, RRID:AB_10915554), CD197 APC-Fire 750 (BioLegend Cat# 353246, RRID:AB_2750147), CD279 PE-Cy7 (BD Biosciences Cat# 561272, RRID:AB_10611585), CD366 BV421 (BioLegend Cat# 345007, RRID:AB_10900073), CD185 BB515 (BD Biosciences Cat# 564624, RRID:AB_2738871), CD95 PE (BioLegend Cat# 305608, RRID:AB_314546), CD27 BV650 (BD Biosciences Cat# 563228, RRID:AB_2744352), and CX3CR1 PE-Dazzle 594 (BioLegend Cat# 341624, RRID:AB_2687152). The samples were incubated at 4°C for 15 minutes, protected from light. Lastly, the samples were incubated in 400ul of 1X BD FACS Lysing Solution (BD 349202) at room temperature for 15 minutes and frozen at −80°C. Thawed samples were collected in batches on a BD Symphony A1 and analyzed using FlowJo (RRID:SCR_008520) v10.8 software.

Cytokines were analyzed in serum on the O-Link platform using the Target 48 panel according to the manufacturer’s instructions. Cryopreserved serum samples were shipped in batch to Psomagen (Rockville, MD) for testing. A total of 45 plasma analytes were tested.

Data Availability

The data generated in the study are available upon request from the corresponding author. Data are not publicly available to protect patient privacy.

Results

Patient characteristics and disposition

Between 2019 and 2022 a total of 47 patients were enrolled to the study. Patients were predominantly male with a median age of 64 (Table 1, Supplementary Table S1). The safety run-in phase included seven patients treated with combination therapy, which included one patient that was replaced who discontinued for rapid tumor progression prior to completing the DLT window. In the next study phase, 19 patients were randomized to combination therapy and 21 to monotherapy. Baseline demographics were similar across treatment arms. The median number of lesions was 4.0 and included metastases to lymph node, lung, liver, and bone. All patients had received prior chemotherapy. Among the 47 enrolled patients at the time of final data lock, five completed the full two years of study treatment, two patients did not receive any study treatment, and 40 discontinued study treatment. The most common reason for treatment discontinuation was disease progression occurring in 30 patients. The mean number of CYT107 doses given was 3.7 in the safety run-in arm and 3.6 in the randomized combination arm.

Table 1.

Patient baseline characteristics

		Safety Run-in: atezolizumab plus CYT107 (n = 7)	Randomized: atezolizumab plus CYT107 (n = 19)	Randomized: atezolizumab monotherapy (n = 21)	Total (n = 47)
Age, median (range) years		59 (46–79)	64 (49–83)	68 (48–76)	64 (46–83)
Sex, male		6 (86)	16 (84)	16 (76)	38 (81)
Race
	White	7 (100)	19 (100)	18 (86)	44 (94)
	Black	0 (0)	0 (0)	3 (14)	3 (6)
Ethnicity (%)
	Hispanic or Latino	0 (0)	0 (0)	2 (10)	2 (4)
	Not Hispanic or Latino	6 (86)	19 (100)	19 (90)	44 (94)
	Unknown	1 (14)	0 (0)	0 (0)	1 (2)
Metastases, median (range), number		5 (3–9)	4 (1–15)	3 (1–11)	4 (1–15)
Site of metastases
	Bone	3 (43)	2 (11)	2 (10)	7 (15)
	Liver	2 (29)	4 (21)	5 (24)	11 (23)
	Lung	4 (57)	5 (26)	8 (38)	17 (36)
	Lymph node	4 (57)	15 (79)	14 (67)	33 (70)
Prior chemotherapy
	Yes	7 (100)	19 (100)	21 (100)	47 (100)
	No	0 (0)	0 (0)	0 (0)	0 (0)
Prior surgery
	Yes	2 (29)	13 (68)	11 (52)	26 (55)
	No	5 (71)	6 (32)	10 (48)	21 (45)
Prior radiation
	Yes	2 (29)	1 (5)	1 (5)	4 (9)
	No	5 (71)	18 (95)	20 (95)	43 (91)

Efficacy

The efficacy analysis set included all 40 randomized patients. The combination of atezolizumab plus CYT107 resulted in an objective response rate (ORR) of 26.3% versus 23.8% for atezolizumab alone and the primary endpoint was not met (p = 0.428) (Figure 1, Supplementary Table S2). The complete response rate was 10.5% for the combination versus 4.8% for the monotherapy arm. The median duration of response was not reached in either group. However, three of the responders in the combination arm (60%) had ongoing responses exceeding 21 months versus one in the monotherapy arm (20%). Clinical benefit rate was comparable between groups at 36.8% for the combination arm vs 42.9% for the monotherapy arm, p = 0.702. The median progression-free survival was 2.1 months (95% CI 1.2, 6.1) in the combination arm versus 2.2 months (95% CI 1.8, 4.6) in the monotherapy arm (p = 0.413) (Supplementary Figure S1). Overall survival was 9.1 months (95% CI 2.6, NC) in the combination arm versus 10.4 months (95% CI 4.7, 20.7) in the monotherapy arm (p = 0.451) (Supplementary Figure S2).

Figure 1.

(A) Summary of objective response rate (ORR) by treatment arm as measured by RECISTv1.1 for all randomized patients. No difference in ORR was observed, though the CR rate was higher in the combination arm. (B) Swimmer plot for duration of response as measured by RECIST v1.1 in all randomized patients.

Safety

The safety analysis set for the CYT107 plus atezolizumab included seven patients in the safety run-in combined with 19 patients randomized to the combination. Treatment with CYT107 was generally well tolerated, with no DLTs observed. No patients receiving CYT107 stopped treatment for unacceptable toxicity. A mean of 3.6 doses of CYT107 were given, and 77% of patients completed all 4 scheduled doses. Any-grade treatment-related adverse event (TRAE) frequency was comparable between groups (Table 2). The most common TRAEs were fatigue, skin rash, diarrhea, nausea, creatinine elevation, and anemia. Although fever, chills, and headache were more common with CYT107 plus atezolizumab, there was an overall numerically lower frequency of grade 3–4 treatment-related adverse events (TRAEs) for the combination compared with atezolizumab alone (21% versus 37%). Similarly, immune-mediated TRAEs were less frequent in the CYT107 plus atezolizumab arm (50% vs 68%). Injection site reactions were observed in only three patients (12%) treated with CYT107, and all events were low grade.

Table 2.

Treatment-related adverse events occurring in >10% of patients or immune-related (safety analysis set)

Adverse Event		Grade	Atezolizumab plus CYT107 (n = 26)	Atezolizumab monotherapy (n = 19)
General, n (%)	Any	Any	21 (81)	14 (74)
		1–2	17 (65)	7 (37)
		3–4	4 (15)	7 (37)
	Anemia	1–2	2 (8)	4 (21)
		3–4	1 (4)	1 (5)
	Diarrhea	1–2	1 (4)	8 (42)
	Constipation	1–2	1 (4)	2 (11)
	Nausea	1–2	5 (19)	3 (16)
	Anorexia	1–2	3 (12)	2 (11)
	Fatigue	1–2	10 (38)	7 (37)
		3–4	1 (4)	1 (5)
	Fever	1–2	6 (23)	1 (5)
	Chills	1–2	5 (19)	0 (0)
	AST increased	1–2	4 (15)	2 (11)
	ALT increased	1–2	4 (15)	1 (5)
	Alkaline phosphatase increased	1–2	4 (15)	1 (5)
	Creatinine increased	1–2	4 (15)	3 (16)
		3–4	0 (0)	1 (5)
	Lymphopenia	1–2	2 (8)	4 (21)
	Thrombocytopenia	1–2	1 (4)	4 (21)
	Leukopenia	1–2	0 (0)	3 (16)
	Hyperglycemia	1–2	0 (0)	3 (16)
	Hypoalbuminemia	1–2	2 (8)	2 (11)
	Hyponatremia	1–2	2 (8)	3 (16)
	Muscle weakness	1–2 3–4	0 (0) 1 (4)	2 (11) 0 (0)
	Headache	1–2	3 (12)	0 (0)
	Injection site reaction	1–2	3 (12)	0 (0)
Immune-related, n (%)	Any	Any	13 (50)	13 (68)
		1–2	13 (50)	10 (53)
		3–4	0 (0)	3 (16)
	Colitis	1–2	0 (0)	0 (0)
		3–4	0 (0)	1 (5)
	Pruritis	1–2	5 (19)	3 (16)
	Rash	1–2	5 (19)	4 (21)
	Hypothyroidism	1–2	0 (0)	1 (5)
	Pancreatitis	1–2	0 (0)	0 (0)
		3–4	0 (0)	1 (5)

Pharmacodynamic activity

CYT107 is a human recombinant IL-7 which is expected to cause activation and proliferation of lymphocytes. Thus, we investigated whether expansion of lymphocyte subsets would be observed in the periphery through serial PBMC analysis. We found that within one week of CTY107 administration both CD4⁺ and CD8⁺ T lymphocyte subsets were expanded significantly in the combination arm but not the control arm (Figure 2A, B) with expansion plateauing around two-fold for both CD4⁺ and CD8⁺ cells at three weeks. Elevated lymphocyte numbers persisted through cycle 4. When analyzing the increases in absolute T cell numbers, we noted that the fold-increase of CD4⁺ T cells was associated (p=0.018) with clinical benefit, but only in the combination arm (Figure 2C). A similar trend of clinical benefit with fold-change of CD8⁺ T cells was observed in the combination arm, though this was not statistically significant (Figure 2D). This clinical benefit was observed in cell counts enumerated in the 21-color whole blood assay analyses (Figure 2C, D) as well as CD4⁺ and CD8⁺ counts enumerated by CLIA testing (analyses not shown).

Figure 2.

(A-B) Fold-change from baseline of T lymphocyte subsets after one dose of CYT107. At 1-week CD4⁺ (p<0.00005) and CD8⁺ (p<0.00005) T lymphocytes were expanded, as determined by CLIA clinical CD4⁺ and CD8⁺ testing. (C-D) The fold-change expansion of lymphocyte subsets was most pronounced in combination arm patients with response to therapy. T lymphocytes were enumerated in whole blood using 21-color flow cytometry. Data are shown for cycle 3 day 1 compared with baseline. (E-F) Responders in the experimental arm showed significant expansion of T_scm cells as analyzed by 14-color flow cytometry on PBMCs, represented as fold-change over baseline of % T_scm within each group by compartment. Expansion was greater in responders vs non-responders. (G-H) PD1 mean fluorescence intensity (MFI) was evaluated and analyzed by 14-color flow cytometry on PBMCs stratified by responders and non-responders. Experimental Arm = CYT107 plus atezolizumab, Control Arm = atezolizumab alone, R = responder, NR = non-responder, data shown as mean +/− SEM.

To investigate whether subsets of CD4⁺ and CD8⁺ T cells might be implicated in the association of expanded T cells with the observed clinical responses, we performed extensive analyses of 21-color (whole blood) and 14-color (PBMC) immunophenotyping data. Whole blood analyses revealed significant expansion of absolute cell counts for both CD4⁺ and CD8⁺ central memory, effector memory, and naïve T cell subsets in the combination arm, but not the control arm (Supplementary Figure S3). Within those subsets, detailed multiparameter 14-color flow cytometry of PBMCs identified the T memory stem cell (T_scm) population to be the most highly expanded subset (Supplementary Figure S4), increasing up to 7-fold (CD4⁺) and 12-fold (CD8⁺) over baseline in patients on combination therapy, but not those on atezolizumab alone (Supplementary Figure S4 and Figure 2E, F). Most strikingly, these cells were expanded to a greater extent in responding patients relative to those patients who did not respond (Figure 2E, F). This expansion peaked between 1–3 weeks after starting therapy. Interestingly, T_scm cells in patients with clinical response were observed to have lower PD1 expression at baseline compared to T_scm cells from patients without a clinical response (Figure 2G, H). The low PD1 expression on these lymphocytes also persisted throughout treatment. Lower CX3CR1 expression on T_scm cells was noted in patients with response to treatment in both groups, although not reaching statistical significance (Supplementary Figure S4).

In addition to assessing the pharmacodynamic activity on circulating PBMC, we also evaluated cytokine profiles using a multiplex assay. Analyzing baseline circulating cytokines for up to 45 analytes, we observed that increased baseline circulating CCL4, and decreased TNF or VEGFA were associated with response to therapy (Figure 3A–C). We also analyzed plasma cytokines at 3 timepoints after treatment (Figure 3D). VEGFA was consistently lower in responders at all timepoints. Other cytokines with consistent differences in at least 2 on-treatment time points included CCL11, which was higher in responders, and CSF1, CXCL8 and MMP1, which were lower in responders. Increased levels over baseline of circulating CXCL9, CXCL10, IL-18, TNF and MMP12 were seen in patients in the combination arm (Supplementary Figure S5). CXCL9, CXCL10 bind CXCR3 and thus act upon T_scm cells¹⁸.

Figure 3.

Plasma samples were obtained and analyzed on the O-Link platform. (A-C) Baseline differences between responders and non-responders were observed for CCL4, VEGF-A and TNF as shown with individual values. (D) Heatmap showing ratio for each cytokine R/NR (Responder/Non-Responder) across multiple timepoints. EOT = End of Treatment.

Discussion

This is the first study, to our knowledge, utilizing an IL-7 agonist immunotherapy in combination with anti-PD-1/L1 blockade in advanced urothelial cancer. The combination of CYT107 at the lowest active dose level and atezolizumab was shown to be well-tolerated and resulted in vigorous systemic lymphocyte expansion. While the primary endpoint of improvement in objective response rate was not met, a higher rate of complete response and greater proportion of durable responses were noted in the combination arm versus atezolizumab monotherapy. Indeed, while the trial size was limited, the doubling of the CR rate with the addition of CYT107 supports the clinical activity of this agent. An unexpected finding was that the rate of immune-related TRAEs and high-grade TRAEs were lower in the combination therapy arm. From this study, several important findings emerged that will be informative to future T cell-augmenting immunotherapy studies in urothelial cancer.

On target CD4⁺ and CD8⁺ T lymphocyte expansion was observed rapidly within seven days and persisted for three months despite CYT107 only being dosed during cycle 1. The expanded population was largest in the T_scm population. T_scm cells are a population of memory lymphocytes that have stem cell-like properties and can maintain self-renewal and polyfunctional activity including effector function¹⁸. In this study, further characterization of T_scm by PD-1 surface expression revealed that patients with response to therapy had low PD-1 expression on T_scm cells. This observation is consistent with preclinical studies that found PD-1 low T_scm cells proliferate more vigorously in response to stimulation with CD3, CD28, and IL-7 compared with memory T lymphocytes expressing PD-1 and TIGIT¹⁹. Additionally, CD4⁺ T_scm cells in responding patients showed a trend towards lower CX3CR1 expression, indicating that a less differentiated T_scm state is beneficial for efficacy²⁰. Of note, there was no difference in baseline CD4⁺ and CD8⁺ T lymphocyte counts between the control arm and experimental arm, or between responders and non-responders.

Nonetheless, this expansion of lymphocytes in our study was not associated with an improved objective response rate compared to atezolizumab alone. This discrepancy may be explained by several factors. The fold expansion of lymphocytes may not have been large enough to induce a clinical response. Because this was the first trial in combination with immune checkpoint blockade, caution was taken in using the lowest active dose of CYT107. Given the tolerability that we have observed, higher and more prolonged dosing of CYT107 beyond cycle 1 may be needed. Additionally, CYT107 at this dose level may be augmenting antigen-specific T lymphocytes that are already responsive to checkpoint blockade alone. A pre-existing T cell-inflamed tumor microenvironment is linked with responsiveness to immune checkpoint blockade, and IL7 agonism may not overcome barriers restricting antigen priming and T cell infiltration into the tumor^4,21. PD-1 low T_scm expansion was most pronounced in patients on the combination arm with clinical response. Taken together, these findings may explain why ORR was similar between arms, but responses were deeper and more durable with combination therapy.

The cytokine profiling in our trial indicated CCL4 was higher in patients with response to therapy. This cytokine is associated with recruitment of tumor antigen cross priming dendritic cells necessary for T-cell activation^4,22–24. In contrast VEGFA and TNF were higher in patients without response, and both may result in immune inhibitory activity via immune suppressive myeloid-lineage cells or modulation of CD8⁺ T cell dysfunction^25,26. Combination treatment also led to increases in CXCL9 and CXCL10, which impact T_scm cells. Thus, the observed cytokine pattern in our study further supports the hypothesis that CYT107 exerts its effect more likely in tumors with a pre-existing T cell-inflamed microenvironment.

Recently, the EV-302 study established the antibody drug conjugate enfortumab vedotin plus the anti-PD-1 checkpoint inhibitor pembrolizumab as a new standard of care for first line advanced urothelial cancer¹². The objective response rate was 68% with a near doubling of the overall survival to 31.5 months when compared to platinum-based chemotherapy. Our study treated patients after platinum therapy, which was where anti-PD-1/L1 immunotherapy was indicated, and most commonly used, prior to the reporting of EV-302. Even with the recent change in first line therapy, the findings in our study remain important given that resistance to anti-PD-1/L1 checkpoint blockade continues to be an unmet need. Only half of patients were alive and without progression at 12 months following enfortumab vedotin and pembrolizumab therapy. Thus, it remains critical to identify and circumvent mechanism of immunotherapy resistance to increase durable responses in urothelial cancer. Furthermore, there is no established standard treatment after this combination, which creates an opportunity for investigating T cell-augmenting therapies that may restore or prolong the immune responses with enfortumab vedotin plus pembrolizumab.

This trial had several limitations. First is that the United States regulatory approval for atezolizumab was withdrawn after atezolizumab failed to confirm an overall survival benefit in combination with front-line chemotherapy. This, in combination with the fact that the standard of care was rapidly changing during the duration of trial accrual, led to slow overall accrual. The original trial design was to randomize 48 patients, but given these factors, and the planned disbandment of the Cancer Immunotherapy Trials Network, accrual was halted early after 40 patients were randomized. Finally, because this was the first study of CYT107 in combination with a checkpoint blocker, the lowest identified active dose level of CYT107 was chosen and only given for one cycle, which may have limited its efficacy.

In summary, we demonstrate that CYT107 can be safely combined with atezolizumab and resulted in a doubling of the CR rate, although the combination did not significantly improve ORR compared with atezolizumab alone. Testing a higher IL-7 dose with repeating cycles to increase efficacy is warranted based on observations described in this study. Correlative studies revealed several insights that will inform future investigation. To improve the potential efficacy for T-cell activating drugs or cellular therapies in urothelial cancer, future clinical trial designs should consider barriers in the tumor microenvironment. Strategies for optimization may include patient selection based on clinical variables (i.e., maintenance in patients with prior response to immune checkpoint blockade), biomarker-based selection (i.e., inclusion of T-cell inflamed tumors), or combination therapy with novel agents that trigger immune priming, promote T lymphocyte trafficking, or inhibit immune suppressive cells.