Combination of denosumab and immune checkpoint inhibition: experience in 29 patients with metastatic melanoma and bone metastases

Abstract

Background

PD-1 inhibition (PD-1i) is the standard of care in melanoma and other malignancies. In patients with bone metastases of solid tumors, the monoclonal antibody denosumab directed against RANKL is approved for the prevention of skeletal-related events. However, RANKL is not only relevant in osteoclastogenesis, but also has immunological effects. Hence, we aimed at investigating, whether the combination of PD-1i and denosumab produces synergistic effects in metastatic melanoma treatment.

Methods

We retrospectively collected and analyzed clinical data of metastatic melanoma patients with bone metastases, who received PD-1i and denosumab therapy.

Results

29 patients were identified with a median age of 60.7 years: 20 were male and 9 were female. 20 patients (69%) were in stage IV M1c and 9 (31%) in stage IV M1d; 52% had an increased serum LDH. 24 patients (83%) received PD-1i as first-line therapy and five patients (17%) as second- or third-line therapy. 13 patients received the triple combination nivolumab, ipilimumab and denosumab (N + I+D), 16 patients received PD-1i and denosumab (PD-1i + D). Within a median follow-up time of 19.8 months, 17 patients progressed with a median time to progression of 6 months. The objective response rate was 54% in the N + I + D group and 50% in the PD-1i + D group. Recalcification of bone metastases was radiologically observed in 18 (62%) patients. No unexpected treatment-related adverse events emerged.

Conclusions

The combination therapy of metastatic melanoma with PD-1i and denosumab was feasible without unexpected safety issues and showed a promising efficacy signal. Further investigation in prospective studies is needed.

Introduction

Immune checkpoint blockade (ICB) targeting the checkpoints CTLA-4 and PD-1 became the standard treatment of advanced melanoma since it has been shown to prolong PFS and OS [1–3].

Bone was found to be the fourth most common site of melanoma visceral metastasis after lung, liver and brain in 11–18% of patients [4]. These bone metastases can lead to skeletal-related events (SREs), such as pathological fracture, spinal cord compression, and severe bone pain [5].

Denosumab, a recombinant fully monoclonal human IgG2 antibody directed against the receptor activator of nuclear factor kappa-B ligand (RANKL), represents a bone-modifying agent and is approved for the prevention of SREs in advanced solid tumors with metastasis to the bone. Denosumab blocks the binding between RANKL and its receptor RANK on osteoclasts resulting in the inhibition of osteoclastogenesis. However, the RANK/RANKL pathway also plays a role in many other processes [6, 7], e.g., immunological processes as it is a member of the tumor necrosis factor (TNF) family starting at thymic development of T cells to regulation of T cell responses in the periphery [8].

Therefore, the combined blockade of immune checkpoints as well as RANKL could have synergistic effects. This has been suggested by a case report of a patient with metastatic melanoma and widespread bone metastases, who showed a remarkable response to the CTLA-4 inhibitor ipilimumab in combination with denosumab [9]. Furthermore, murine tumor models suggest synergistic effects of the combined administration of CTLA-4 inhibitors and denosumab [10] or PD-1 inhibition (PD-1i) and denosumab [11].

To obtain further evidence for the clinical tolerability and efficacy of checkpoint inhibition, we retrospectively collected data from patients treated at four German skin cancer centers with denosumab and PD-1i for metastatic melanoma and bone metastases.

Methods

Study design

This is a retrospective multicenter study initiated by the German Dermatologic Cooperative Oncology Group (DeCOG). Patients meeting the following criteria were eligible: diagnosis of unresectable metastatic melanoma including bone metastases, therapy with denosumab and PD-1i with nivolumab, pembrolizumab or nivolumab plus ipilimumab combination therapy with a minimum concurrent administration of 4 weeks. These treatments were performed as the standard of care according to the summary of product characteristics (SmPC), e.g., before denosumab treatment a dental check was performed and calcium was substituted during therapy.

Data collection

Participating centers retrospectively searched their electronic databases for eligible patients. Data were provided on a standardized form including demographics (age, sex), melanoma characteristics (preceding therapies, stage of disease according to the AJCC classification [12], ECOG performance status), the PD-1i under investigation (type of inhibitor, date of therapy start, best response according to RECIST 1.1 criteria [13] [graded into complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD)], change of bone lesions such as sclerosis and details on the denosumab therapy (onset, duration). Imaging for restaging was performed every 2–3 months according to German melanoma guidelines [14]. Adverse events were graded according to Common Terminology Criteria for Adverse Events (CTCAE) version 4.03 [15].

Statistical analysis

The Kaplan–Meier method was used to calculate estimates of PFS und OS using the software XLSTAT 2018. OS and PFS were defined as the time between the first administration of the immune checkpoint inhibitor and death or occurrence of disease progression, respectively. The results of OS and PFS were both presented in months.

Results

Patients

29 eligible patients were identified from four clinical centers. 20 patients were male, 9 were female, with a mean age of 60.7 years (range 35–82 years, Table 1). 13 of the patients received a combination therapy with nivolumab, ipilimumab and denosumab (N + I+D), while 16 others received a monotherapy with a PD-1-inhibitor plus denosumab (PD-1i + D). Eight patients received nivolumab and eight patients received pembrolizumab, respectively. The minimum of concomitant ICB and denosumab exposure was 1 month, the median 7 months and the maximum 44 months.

Table 1.

Baseline characteristics of patients receiving checkpoint inhibitors plus denosumab treatment

	Nivolumab + ipilimumab + denosumab (N = 13)	PD-1-inhibitor + denosumab (N = 16)	Total (N = 29)
Age (years)
Mean	59.9	61.4	60.7
Range	39–82	35–82	35–82
Sex: no. (%)
Male	10 (77%)	10 (63%)	20 (69%)
Female	3 (23%)	6 (37%)	9 (31%)
Metastasis stage IV: no. (%)
M1c	8 (62%)	12 (75%)	20 (69%)
M1d	5 (38%)	4 (25%)	9 (31%)
Lactate dehydrogenase: no. (%)
Normal	7 (54%)	7 (44%)	14 (48%)
Elevated	6 (46%)	9 (56%)	15 (52%)
BRAF-status: no. (%)
Mutated	10 (77%)	7 (44%)	17 (59%)
Wild type	3 (23%)	9 (56%)	12 (41%)
Therapy line: no. (%)
1st	12 (92%)	12 (75%)	24 (83%)
2nd	1 (8%)	1 (6%)	2 (7%)
3rd	0 (0%)	3 (19%)	3 (10%)

A total of 31% of the patients had stage IV M1d and 69% had stage IV M1c metastatic melanoma, respectively. 52% of the patients had an elevated lactate dehydrogenase (LDH) serum level, and 59% had a BRAF V600 mutation. Four patients had received a BRAF-directed targeted therapy before ICB and denosumab. 24 patients had received PD-1i as the first-line treatment, two as the second-line and three others as the third-line treatment, respectively (Table 1).

Efficacy

In the N + I + D group, a CR was observed in two patients (15%), a PR in five patients (39%), three patients (23%) had a SD and three patients (23%) had a PD. In the PD-1i + D group, three patients (19%) experienced a CR, five patients (31%) a PR, four patients (25%) had a SD and four patients (25%) had a PD (Table 2).

Table 2.

Efficacy of checkpoint inhibitors plus denosumab in 29 patients

	Nivolumab + ipilimumab + denosumab (N = 13)	PD-1-inhibitor + denosumab (N = 16)	Total (N = 29)
Best overall response: no. (%)
Complete response	2 (15%)	3 (19%)	5 (17%)
Partial response	5 (39%)	5 (31%)	10 (35%)
Stable disease	3 (23%)	4 (25%)	7 (24%)
Progressive disease	3 (23%)	4 (25%)	7 (24%)
Objective response: no. (%)	7 (54%)	8 (50%)
Median duration of response (months)	10.2	12.5
Median (range) PFS (months)	6 (0–24)	6 (1–44)
Median (range) OS (months)	NR	NR
Sclerosis of bone metastases: no. (%)	6 (46%)	12 (75%)	18 (62%)

At the time of analysis, progressive disease was observed in 16 patients (8/13 from N + I + D group and 8/16 from PD-1i + D group). Five (17%) patients had died due to tumor progression (1/13 from N + I+D group and 4/16 from PD-1i + D group). The median values of OS are not reached (Fig. 1).

Fig. 1.

Kaplan–Meier curves of PFS (a) and OS (b), separated for patients receiving N + I+D and patients receiving PD-1i + D

Bone metastases responding to treatment as assessed radiographically by sclerotic changes within a metastatic deposit were seen in 18 patients (6/13 patients in the N + I + D group; 12/16 patients in the PD-1i + D group) (Fig. 2). All of these 18 patients with sclerotic changes in the bone metastases responded. Of the 11 other patients without sclerotic changes of bone metastases, two patients have reached PR, two patients had a SD and seven patients had a PD.

Fig. 2.

Sclerotic changes within bone metastases as response to treatment. The left CT scan shows the metastases before treatment, the right CT scan with sclerotic changes of the bone metastases after 12 months of treatment with N + I+D. The patient achieved a partial response

The overall response rate in the subgroup of patients with BRAF V600 mutation melanoma (47%; 1 CR and 7 PR) was similar compared to patients with BRAF V600 wild type (58%; 4 CR and 3 PR) (Table 3). Two (13%) patients with elevated LDH had a CR, seven (47%) patients had a PR, two (13%) had a SD and four (27%) patients had a PD. In the group of patients with normal LDH, CR was observed in three patients (21%), PR in three patients (21%), three patients (21%) had a SD and five patients (36%) had a PD. Overall, the patients with an elevated LDH level had a slightly higher response rate (60%) compared to patients with normal levels of LDH (43%). Four patients (20%) without brain metastases (stage IV M1c) achieved a CR, seven patients (35%) had a PR, seven patients (35%) had a SD and three patients (15%) had a PD. In the group of patients in stage IV M1d, a CR was observed in one patient (11%), a PR in three patients (33%), a PD in five patients (56%) and no patients in this group had a SD. Thus, response rates of patients with stage IV M1d were 44% and those with stage IV M1c were 55% (Table 3).

Table 3.

Efficacy of checkpoint-inhibitor plus denosumab treatment in certain subgroups

	BRAF V600E-mutation (N = 17)	BRAF V600E wild type (N = 12)	Total (N = 29)
Best overall response: no. (%)
Complete response	1 (6%)	4 (33%)	5 (17%)
Partial response	7 (41%)	3 (25%)	10 (34%)
Stable disease	5 (29%)	2 (17%)	7 (24%)
Progressive disease	4 (24%)	3 (25%)	7 (24%)
Objective response: no. (%)	8 (47%)	7 (58%)
	Elevated LDH level (N = 15)	Normal LDH level (N = 14)	Total (N = 29)
Best overall response: no. (%)
Complete response	2 (13%)	3 (21%)	5 (17%)
Partial response	7 (47%)	3 (21%)	10 (34%)
Stable disease	2 (13%)	5 (36%)	7 (24%)
Progressive disease	4 (27%)	3 (21%)	7 (24%)
Objective response: no. (%)	9 (60%)	6 (43%)
	M1c (N = 20)	M1d (N = 9)	Total (N = 29)
Best overall response: no. (%)
Complete response	4 (20%)	1(11%)	5 (17%)
Partial response	7 (35%)	3 (33%)	10 (34%)
Stable disease	7 (35%)	0 (0%)	7 (24%)
Progressive disease	2 (10%)	5 (56%)	7 (24%)
Objective response: no. (%)	11 (55%)	4 (44%)

Adverse events

Treatment-related adverse events of any grade occurred in 85% of the patients in the N + I + D group and 50% of those in the PD-1i + D group (Table 4). The most frequently reported treatment-related adverse events in N + I+D group were colitis of grades 3–4 (in 31% of the patients), followed by hypothyroidism of grade 1 or 2 (in 23%). Fatigue, rash and dyspnea were observed in two patients each (15%). On the other hand, fatigue and rash of grade 1 or 2 (in 25% of patients) were the most common adverse events in the PD-1i + D group, and no grade 3 or 4 treatment-related adverse events were reported in this group. Two patients (15%) had to discontinue the treatment due to adverse events in the N + I + D group, but none in the PD-1i + D group. The discontinuations were due to colitis (grade 4) and hepatitis (grade 4), respectively. Adverse events after injection of denosumab were reported in two patients. Myalgia occurred as AE in one patient, whereas the other had pyrexia. The AEs occurred repetitively within 1–2 days after denosumab injection but were not observed after ICB-application and were, therefore, considered as denosumab-related AEs. No other adverse events under denosumab therapy including osteonecrosis of the jaw were reported in our patients.

Table 4.

Adverse events in patients receiving checkpoint inhibitors plus denosumab treatment

	Nivolumab + ipilimumab + denosumab (N = 13)		PD-1-ihibitor + denosumab (N = 16)
	Grade 1–2	Grade 3–4	Grade 1–2	Grade 3–4
Number of patients (%)
Treatment-related adverse events	11 (85%)		8 (50%)
Fatigue	2 (15%)	0	4 (25. %)	0
Pruritus	1 (8%)	0	1 (6%)	0
Vitiligo	0	0	2 (12%)	0
Rash	2 (15%)	0	4 (25%)	0
Nausea/other gastrointestinal symptoms	1 (8%)	2 (15%)	2 (12%)	0
Decreased appetite	0	1 (8%)	1 (6%)	0
Increased aspartate or alanine transaminase	0	1 (8%)	0	0
Increased lipase and amylase	1 (8%)	0	0	0
Hypothyroidism	3 (23%)	0	1 (6%)	0
Colitis	0	4 (31%)	0	0
Arthalgia	1 (8%)	0	2 (12%)	0
Dyspnea	2 (15%)	0	3 (19%)	0
Pyrexia	0	1 (8%)	0	0
Myasthenic syndrome	0	1 (8%)	0	0
Treatment-related adverse events leading to discontinuation		2 (15%)		0

The severity of adverse events was graded according to Common Terminology Criteria for Adverse Events (CTCAE) version 4.03 (for download see link http://evs.nci.nih.gov/ftp1/CTCAE/CTCAE_4.03_2010-06-14_QuickReference_8.5x11.pdf)

Discussion

This study aimed to provide clinical evidence that the use of denosumab as a targeted therapy against RANKL in combination with immune checkpoint blockade is feasible and may enhance the antitumor effect in metastatic melanoma patients with bone metastases.

We identified 29 patients, who received PD-1-directed checkpoint inhibition plus denosumab for metastatic melanoma with bone metastases.

With regard to AEs, the overall incidence of treatment-related AEs was significantly higher in the N + D+I group (85%) than in the PD-1i + D group (50%). The incidence of treatment-related AEs of grade 3 or 4 was only reported in N + I + D, notably for colitis. Overall, the safety profile of the combination therapy remained consistent with previous experiences within the combination therapy with nivolumab and ipilimumab and with PD-1i monotherapy [16]. For denosumab, a well-known serious adverse event is osteonecrosis of the jaw which occurs in 1–2% of patients [17]. In our case series, we observed no osteonecrosis of the jaw. Thus, our case series did not show any unexpected safety issues with regard to type, frequency or severity of AEs.

With regard to efficacy, response rates of the patients in the N + I + D group (54%) were slightly higher compared to the response rates of the patients in the PD-1i + D group (50%). These response rates are in line with previous studies investigating ICB in melanoma without concomitant denosumab treatment. The Checkmate 67 study revealed a response rate of 45% for nivolumab and 58% for nivolumab plus ipilimumab [18]. Furthermore, the Keynote 6 study showed a response rate of 42% for pembrolizumab [16]. To our knowledge, there are no data focusing on the response of melanoma bone metastases on ICB or denosumab alone. In our experience, bone metastases also respond to the treatment if extraosseous metastases respond. This is also reflected in our study, where all 18 patients with sclerotic changes of bone metastases had an overall response. Out of the 11 patients without sclerotic changes, there were only two overall responders.

Recently, there were two case series reporting the combination of ICB and denosumab. The study from Liede et al. [19] retrieved melanoma and lung cancer patients from the Flatiron Health electronic health records, who started both ICB (CTLA-4 inhibitor ipilimumab or PD-1 inhibitor nivolumab or pembrolizumab) and denosumab within a period of 30 days. This analysis reported a 47% response rate for PD1i + D treatment in 30 patients, which is similar to the response rate of 50% observed in our patients. However, they did not report on the combination of N + I+D or adverse events. Furthermore, Afzal and Shirai [20] performed a retrospective chart review of melanoma patients from two cancer centers who received ICB in combination with denosumab. Of the 11 patients they identified, 8 received pembrolizumab, 1 nivolumab, 1 ipilimumab, and 1 nivolumab + ipilimumab, respectively. They reported a response in 6/11 patients (55%). Hypocalcemia (2/11 patient grade 2) and myalgia/back pain (5/11 patients grade 1 or 2) were more common in the denosumab group as compared to a matched control group with ICB alone (26 patients with metastatic melanoma without bone metastases).

The mechanism as to how RANKL blockade with denosumab might affect immunotherapy is a matter of speculation. One option is a direct effect on tumor cells. RANK can be expressed by melanoma cells, but RANK expressing melanomas and RANK-negative melanomas have similar rates of bone metastases; thus, the functional role of RANK on melanoma cells is unclear [21–24]. Another option is an effect of the RANKL/RANK pathway in immune reactions. RANKL expression has been described on various immune cells, such as regulatory T cells, myeloid suppressor cells and CD8 + as well as CD4 + T cells [25–29]. Thus, immunological effects of denosumab are possible, which is also suggested by murine models, where the concomitant administration of anti-CTLA-4 and RANKL-Inhibitor resulted in an increased number of tumor-infiltrating CD8 + T cells [10, 30].

Of course, our case series implies the usual limitations of all retrospective studies, such as possible bias by patient selection and incomplete reporting of data. However, it provides further clinical evidence for the tolerability and efficacy of both N + I+D as well as PD-1i + D. These findings encourage the conduct of prospective clinical trials investigating the combination of PD-1i plus denosumab.

Appropriate prospective clinical trials are currently ongoing. The CHARLI-trial (NCT 03161756) is a two-arm study performed in Australia in patients with metastatic melanoma. Arm A investigates denosumab plus nivolumab, arm B denosumab plus nivolumab + ipilimumab. The primary endpoints are median PFS and the occurrence of grade 3–4 selected immune-related adverse events of interest. The BONEMET trial (EudraCT-Nr. 2016-001925-15) is performed in Germany and includes patients who receive treatment with PD-1i plus denosumab for metastatic melanoma with bone metastases.

The results of these trials are expected earliest in 2021. Since the combination of denosumab plus PD-1-directed checkpoint inhibition is currently frequently used in the clinical care of metastatic melanoma patients, our retrospectively collected evidence is important as a basis for the potential use of this combination.

Conclusion

We show in a retrospective analysis of 29 patients that the combination of PD-1i or nivolumab plus ipilimumab with denosumab is feasible with no unexpected safety concerns and promising efficacy. Thus, prospective trials are encouraged to further investigate this combination.

Abbreviations

AEs

Adverse events

AJCC

American Joint Committee on Cancer

CR

Complete response

CTCAE

Common Terminology Criteria for Adverse Events

CTLA4

Cytotoxic T-lymphocyte-associated protein 4

EMA

European Medicines Agency

FDA

Food and Drug Administration

ICB

Immune checkpoint blockade

LDH

Lactate dehydrogenase

N

Number

N + I+D

Nivolumab plus ipilimumab plus denosumab

NR

Not reached

ORR

Objective rate response

PD

Progressive disease

PD-1i

Programmed cell death protein1 inhibitor/inhibitors

PD-1i + D

PD1 inhibitor plus denosumab

PR

Partial response

RANK

Receptor activator of NF-kB

RANKL

Receptor activator of NF-kB ligand

RECIST

Response evaluation criteria in solid tumors

SD

Stable disease

SmPC

Summary of product characteristics

SREs

Skeletal-related events

Author contributions

Authors whose names appear on this paper have contributed sufficiently to the scientific work and, therefore, share collective responsibility and accountability for the results. YA and RG planned the project. All authors contributed to the data collection. YA and RG contributed to the interpretation of the results. YA took the lead in writing the manuscript. All authors provided critical feedback and helped to shape the manuscript.

Funding

No specific funding was received for this study.

Compliance with ethical standards

Ethics approval

This is a retrospective multicenter study initiated by the German Dermatologic Cooperative Oncology Group (DeCOG). Human participants (melanoma patients) were involved but only by the retrospective chart review. Ethics approval for the retrospective data collection and anonymous analysis without informed consent of the patients was obtained from the Ethics committee of Hannover Medical School (vote number 1612). This vote states that there are no ethical concerns with regard to the anonymous retrospective analysis of routine clinical care data from own patients and that there is no general need to notify the ethics committee on such studies. The patient who is shown in Fig. 2 provided written informed consent for the use of his CT scans within this publication.

Conflict of interest

Ralf Gutzmer served as speaker to Roche, Bristol-Myers Squibb (BMS), Novartis, Merck Serono, Merck Sharp & Dohme (MSD), Almirall-Hermal, Amgen, Merck Serono, Pierre Fabre, SUN Pharma and received research grants from Novartis, Pfizer, Johnson & Johnson. He served as consultant to Roche Pharma, BMS, Novartis, MSD, Almirall-Hermal, LEO-Pharma, Amgen, Pfizer, Pierre Fabre, Roche Posay, Merck Serono, Regeneron, SUN Pharma and Incyte. Selma Ugurel received research support from BMS, medac, and Merck Serono and serves as speaker/consultant to BMS, MSD, Merck Serono, and Roche, and received travel grants from BMS, Medac, and MSD. Jürgen C. Becker received speaker fees from Amgen, MerckSerono, and Pfizer, advisory board honoraria from Amgen, CureVac, eTheRNA, Lytix, MerckSerono, Novartis, Rigontec, and Takeda as well as research grants from Alcedis, Boehringer Ingelheim, BMS and Merck Serono. He also received travel fees from 4SC and Incyte. Carsten Weishaupt served as speaker to Amgen, BMS, Curevac, La Roche Posay, Leo Pharma, MSD, Novartis, Roche, Takeda, TEVA. All other authors declared that they have no conflict of interest.