Targeted therapies and PARPi therapy response following ICI therapy failure in advanced melanoma: a case series

Abstract

Poly (ADP-ribose) polymerase inhibitors (PARPi) have shown efficacy in treating cancers with homologous recombination deficiency (HRD), including subsets of melanoma. However, the potential synergy between PARPi and standard melanoma therapies remains understudied. Here, we report two cases of advanced metastatic melanoma refractory to standard-of-care treatment that demonstrated durable partial responses following the addition of PARPi in combination with immune checkpoint inhibitors (ICIs) and BRAF/MEK inhibitors. Both patients exhibited homologous recombination repair (HRR) pathway mutations and tolerated the combinatory regimens well, achieving progression-free survival of more than 11 months. Mechanistically, PARPi may enhance immunogenicity to ICI therapy via activation of the cyclic GMP-AMP synthase-stimulator of interferon (cGAS-STING) pathway and modulation of programmed death ligand 1 (PD-L1) expression. Preclinical studies also support synergism between PARPi and BRAF/MEK-targeted therapies. This report highlights the potential for PARPi to be integrated into advanced melanoma treatment, particularly in HRD tumors and in combination with ICI and targeted therapies. Although limited by the small sample size, our findings support the rationale for ongoing clinical trials evaluating PARPi-based combinations and underscore the need for further studies to clarify the optimal sequencing and combinations of therapy.

HIGHLIGHTS

1. PARP inhibition (PARPi) offers a mechanistically rational option for advanced melanoma with homologous recombination deficiency (HRD).

2. Two heavily pretreated metastatic melanoma patients achieved durable partial responses after PARPi-based combination therapy.

3. Durable disease control was achieved with PARPi plus immune checkpoint inhibitors (ICI), and with triple therapy (PARPi + ICI + BRAF/MEK inhibition).

4. Mechanistic data support synergy between PARPi and ICI.

1. Introduction

Poly (ADP-Ribose) Polymerase inhibitors (PARPi) have demonstrated great efficacy as a treatment option for advanced cancer patients with homologous recombination deficiency (HRD). Mechanistically, the treatment works by allowing the accumulation of double-stranded DNA (dsDNA) breaks, which are normally repaired through homologous recombination mechanisms in healthy cells. However, in HRD cells, these mechanisms are absent; therefore, these cells experience cell death through synthetic lethality. There are multiple reports outlining positive responses to PARPi in patients with melanoma harboring homologous recombination repair (HRR) mutations [1–3]. However, this favorable response was not seen in patients without HRR mutations in melanoma [1, 2, 4].

Outside of its usage in HRD melanoma, little is known about the potential synergism that may exist in combining PARPi with other standard therapies such as immune checkpoint inhibitors (ICI) and BRAF inhibitors. Additionally, much research is needed to understand how PARPi could fit into the treatment algorithm against advanced melanoma. Here, we report two advanced metastatic melanoma patients who demonstrated a sustained partial response to the combination of PARPi and ICI therapy and PARPi, ICI, and BRAF targeted therapies following disease progression on standard-of-care treatment.

2. Case description

We identified two patients at Siteman Cancer Center of Washington University School of Medicine who presented for metastatic melanoma (MM) management. Both demonstrated sustained partial response (PR) with the combination of PARPi and targeted (BRAFi) and/or immunologic (programmed death ligand 1 (PD-L1) inhibitor) therapies after experiencing progressive disease (PD) on standard-of-care treatment. Treatments were well-tolerated. Patient demographics and clinical data are reported in Table 1.

Table 1.

Patient demographics and clinical data.

	Patient 1	Patient 2
Demographics	65-year-old, Male, Caucasian (non-Hispanic)	60-year-old, Male, Caucasian (non-Hispanic)
PMHx	IDA, colon polyps, arthritis	Treated right anterior abdominal melanoma 30 years prior
Family History	No melanoma or non-melanoma skin cancers	No melanoma or non-melanoma skin cancers
Melanoma Location and Pathology	Primary: unknown Metastasis: left parietal brain lesion	Primary: unknown Metastasis: right axillary lymph node
Stage at diagnosis	Stage IV	Stage IV
Initial Therapy	Left craniotomy with resection of left parietal brain lesion and gamma knife (fSRS 30Gy/5fx) to left parietal lobe	Excisional biopsy of right axillary lymph node
Therapy 2	Ipilimumab (3mg/kg)/nivolumab (1mg/kg) every 3 weeks Target: PD-L 1 Duration: 3 months (4 cycles) Response: n/a AE: Grade 3 pneumonitis	Phase IB of NEO-PV-01 plus Nivolumab clinical trial Target: PD-L 1 Duration: 21.9 months Response: PD AE: Autoimmune hypothyroidism and started on levothyroxine, vitiligo
Therapy 3	Gamma knife to right posterior parietal lobe lesion	Right axillary lymph node dissection at levels 1–3
Therapy 4	Olaparib Target: PARP Duration: 17.5 months Response: PD AE: n/a Interval therapies: SBRT to left cingulate gyrus, Laparoscopic partial hepatectomy of segment 4b with microinvasive ablation of segment 2 liver metastasis, LITT to right parietal lesion, Radiation therapy (fSRS 30 Gy/5fx) to left parietal resection cavity and scalp metastases	Ipilimumab/nivolumab then nivolumab 480mg every 4 weeks only Target: PD-L1 Duration: Ipilimumab/nivolumab for 2.8 months then nivolumab only for 10.4months (28 cycles Nivolumab total) Response: PR AE: vitiligo
Therapy 5	Nivolumab/Relatlimab Target: PD-L1 Duration: 1.9 months Response: PD AE: n/a	Observation Duration: 9.5 months Response: PD
Therapy 6	Nivolumab/Relatlimab + Olaparib Target: PD-L1 and PARP Duration: 11.3 months at last follow-up; therapy ongoing Response: PR and stable disease AE: n/a DFP: 11.5 months	Atezolizumab/NT-17 study Target: PD-L1 Duration: 1 cycle Response: PR and stable disease AE: hospitalized for IO-induced adrenal insufficiency, skin reaction, fatigue, and diarrhea
Therapy 7	n/a	Full dose encorafenib/binimetinib then dose reduced encorafenib 300mg q day and binimetinib 30mg BID Target: BRAF/MEK Duration: full dose for a month then reduced dose for 9.4 months Response: PD AE: hospitalization for fever with undifferentiated shock and skin rash while on full dose. No AE on the reduced dose.
Therapy 8	n/a	Nivolumab/Relatlimab Target: PD-L1 Duration: 6 months Response: PD AE: n/a
Therapy 9	n/a	Nivolumab/Relatlimab + dose reduced encorafenib/binimetinib Target: PD-L1 + BRAF/MEK Duration: 0.7 months Response: PD AE: n/a DFP: 1.06 months
Therapy 10	n/a	Nivolumab/Relatlimab + dose reduced encorafenib/binimetinib + Olaparib Target: PD-L1 + BRAF/MEK + PARP Duration: 15.1 months at last follow-up; therapy ongoing Response: PR and stable disease AE: n/a
Overall Survival	43 months	89.9 months

PR = partial response, CR = complete response, NOS = not otherwise specified, PD = progressive disease, MR = mixed response, LITT = laser interstitial thermal therapy, SBRT = short beam radiation therapy, n/a = not applicable, PD-L1= programmed death ligand 1, BID = twice daily, AE = adverse effects, DFP = disease free progression.

2.1. Patient 1

A 65-year-old Caucasian male with a history of colon polyps presented with speech, calculation, and spatial orientation difficulties. Initial imaging revealed a 2.4 × 2.8 cm left parietal lobe mass with surrounding vasogenic edema. No primary site, lymph nodes (LN) or other metastases were detected on computed tomography (CT) scans or positron emission tomography (PET CT). He had tumor resection via left parietal craniotomy with pathology confirmed MM. His stage 4 MM of unknown primary origin was BRAF wild type, KIT negative, PD-L1 negative, microsatellite instability (MSI) stable and undetected homologous recombination deficiency (HRD) with low loss of heterozygosity (LOH) (28.8%), as determined by Tempus testing. Genetic testing revealed pathogenic germline MUTYH c.1187G > A p.G396D, somatic NRAS c.181C > A p.Q61K, TERT c.-146C > T, and ARID2 c.1468C > T p.Q490* mutations (Table 2). Following gamma knife fractionated stereotactic radiosurgery (fSRS) therapy at 30 Gy in 5 fractions, he commenced ipilimumab (3 mg/kg) and nivolumab (1 mg/kg) every 3 weeks for 4 cycles; however, it was discontinued due to grade 3 pneumonitis. Four months later, a repeat brain MRI showed a new 1 cm falx-adjacent right posterior parietal lesion which was treated with gamma knife therapy. Unfortunately, he had PD 2 months later with increased uptake in the left and right hepatic hemispheres on PET CT, and a new nodular enhancing lesion in the left cingulate gyrus on MRI brain. Standard olaparib dose of 300 mg twice daily (BID) was initiated and continued for 17.5 months. During this time, he also received stereotactic body radiation therapy (SBRT) to the left cingulate gyrus, laparoscopic partial hepatectomy of segment 4b, microinvasive ablation of segment 2, laser interstitial thermal therapy (LITT) to a right parietal lobe, and fSRS therapy at 30 Gy in 5 fractions to the left parietal/scalp metastases. Olaparib was continued until PD on Fludeoxyglucose F18 (FDG) PET/CT showing hypermetabolic lesions in left scalp, left trapezius, tongue base, hepatic segment 7, right hemi-thyroid, right cervical chain and right para-tracheal LNs. The liver lesion showed similar genetic testing to the initial brain lesion with new MAP2K2 (MEK2) p.V131L c.391G > C (variant of uncertain significance (VUS)) and NTRK1 p.E697K c.2089G > A VUS mutations (Table 2). As opposed to the initial brain lesion, this metastasis was PD-L1 positive and LOH at 23% as determined by commercially available next-generation sequencing. Right middle thyroid lobe biopsy also showed MM. Due to diffuse metastases, he started nivolumab/relatlimab but brain MRI showed progression with increased size and vasogenic edema of the parietal resection cavity lesion and bilateral cingulate gyri lesions. One of the cingulate gyri lesions was outside the radiation field, discounting the possibility of pseudoprogression. Olaparib was added and continued until the present day. After 11.3 months of treatment, he exhibited a PR and overall stable disease (SD) (Figure 1). This therapy was generally well tolerated though cycle 15 of nivolumab/relatlimab (at last follow-up) was held due to concerns of organizing pneumonia. At his last follow-up, he had an overall survival of 43 months and 11.5 months progression free.

Table 2.

Melanoma tumor genetic and molecular profiling.

Case	Patient 1		Patient 2
Site	Left parietal brain lesion	Liver	Right axillary lymph node	Right supraclavicular lymph node	Right shoulder
IHC	Negative: BRAF and KIT	Negative: BRAF	Positive: BRAF V600E mutation; VAF 13%	Positive: BRAF V6000E (1+,100%), MLH1 (2+, 100%), MSH2 (2+, 100%), MSH6 (1+, 90%), PMS2 (2+, 100%)	Positive: BRAF V6000E mutation
PD-L1	Negative	Negative	Not tested	PD-L1 SP142 Positive (2+, 50%)	PD-L1 SP142 Positive (1+, 35%)
Mutations	Pathogenic Variant: 1. Germline MUTYH c.1187G > A p.G396D 2. NRAS c.181C > A p.Q61K on exon 3; VAF 40.6% 3. TERT c.-146C > T; VAF 36.5% 4. ARID2 c.1468C > T p.Q490*; VAF 85.1%	Pathogenic Variant: 1. Germline MUTYH p.G396D c.1187G > A on exon 13; VAF 48% 2. ARID2 p.Q490* on exon 11 3. NRAS p.Q61K, exon 3 4. TERT promoter c.-146C > T VUS: 1. MAP2K2 (MEK2) p.V131L c.391G > C on exon 3; VAF 26% 2. NTRK1 p.E697K c.2089G > A on exon 16; VAF 30%	Pathogenic Variant: 1. MAP2K1 (MEK1) p.C121G; VAF 5% 2. TERT c.-146C > T; VAF 10%	Not tested	Pathogenic Variant: 1. BRAF p.V600E c.1799T > A on exon 15; VAF 45% 2. BRCA2 p.F2560fs c.7679_7680delTT on exon 16; VAF 13% 3. TERT promoter c.-146C > T; VAF 24% 4. B2M pathogenic variant p.S14fs c.41_44delCTCT; VAF 15% VUS: 1. MAP2K1 (MEK1) p.C121G c.36TT > G on exon 3; VAF 11% 2. NF1 p.I1605V c.4813A > G on exon 36; VAF 50%
MSI	stable	stable	stable	stable	stable
TMB (mutations per Mb)	High (22.6m/Mb)	High (22m/Mb)	Indeterminate (7mut/Mb)	n/a	High (13 mut/Mb)
HRD	Not detected	n/a	n/a	n/a	n/a
LOH	Low (28.8%)	High (23%)	n/a	n/a	Low (6%)

TMB = tumor mutational burden, MSI = microsatellite instability, MMR = mismatch repair, LOH = loss of heterozygosity, PD-L1 = Programmed death ligand 1, VAF = Variant allele frequency, Mb = megabase, VUS = variant of uncertain significance.

Figure 1.

Patient 1- (A-B): CT-imaging showing partial response to therapy after 3 months of olaparib monotherapy with an anatomic and metabolic improvement of hepatic segment 4 A and 5 lesions [green arrows]. (C-E): MRI brain imaging showing a cingulate gyrus [yellow] and left parietal lesion [orange] when nivolumab and relatlimab treatment started (C), disease progression after 3 cycles of nivolumab and relatlimab treatment (D) and partial response was seen 2 months after olaparib was added to nivolumab and relatlimab and continued beyond 11 months (E).

2.2. Patient 2

A 60-year-old Caucasian male presented with a softball-sized right axillary mass. Initial CT chest and PET showed multiple bilateral lung nodules, enlarged right-sided supraclavicular, subpectoral and axillary LNs, and an indeterminate 2.3 cm right hepatic lobe lesion but no evidence of intracranial disease. Right axillary LN excisional biopsy revealed BRAF V600-positive and MSI stable MM with MAP2K1 (MEK1) C121G and TERT c.-146C > T mutations (Table 2). He was enrolled in phase IB of NEO-PV-01 plus nivolumab clinical trial for treatment of his stage 4 MM of unknown primary (NCT02897765) [5]. He had SD at 5 months but experienced PD at 23.7 months on therapy, with an interval increase in right axillary lymphadenopathy size. He developed vitiligo and immunotherapy (IO)-related hypothyroidism treated with levothyroxine. The right axillary LN was positive for BRAF V600E, MLH1, MSH2, MSH6, PMS2 and PD-L1 SP142 mutations. Imaging 5.5 months after right axillary LN level 1–3 dissection showed PD with interval increase in lymphadenopathy (supraclavicular, periaortic, and left iliac LNs). Ipilimumab/nivolumab was initiated, with SD noted after 2.8 months of therapy. He was transitioned to nivolumab 480 mg every 4 weeks but experienced fatigue and developed IO-induced adrenal insufficiency treated with prednisone. After 28 cycles of nivolumab and SD on imaging, he was placed on observation. PD was noted after 10 months with multiple new hypermetabolic LNs and increased size and uptake of the preexisting LNs above and below the diaphragm on PET CT. A new right middle lobe pulmonary nodule with faint uptake was also suspicious for metastatic disease. Biopsy of a lesion on his right posterior shoulder also showed MM. He was then enrolled in the Atezolizumab/NT-17 study but only received one dose as he was hospitalized for IO-induced adrenal insufficiency (NCT04332653) [6]. As a result, he began fully dosed encorafenib/binimetinib but only tolerated one month of therapy before requiring hospitalization for fever and undifferentiated shock. He was continued on dose-reduced encorafenib/binimetinib (300 mg q day and 30 mg BID respectively), after adrenal insufficiency was ruled out, which was better tolerated during 9.4 months of therapy but had PD after 9.7 months with marked interval increase in size and uptake of the hypermetabolic LNs and the right middle lobe pulmonary nodule. He then received 6 months of nivolumab/relatlimab and exhibited a mixed response after approximately 3 months (with increased subcutaneous nodules and some improvement in abdominal adenopathy), thus dose reduced encorafenib/binimetinib was reintroduced to his therapy. Due to limited response after approximately 1 month of combination nivolumab/relatlimab and encorafenib/binimetinib, standard dose of olaparib 300 mg BID was added. At the time of last follow up (89.9 months), he had completed a total of 24 cycles of nivolumab/relatlimab and has tolerated his triple therapy well. PET CT imaging at this time shows overall SD without new lesions with a disease-free progression of 1.06 months and overall survival of 89.9 months (Figure 2).

Figure 2.

Patient 2- (F-I): CT-imaging showing intensely hypermetabolic right posterior shoulder subcutaneous lesion [blue arrow] before [F & G] and 15 months after [H&I] the addition of PARPi therapy.

3. Discussion

This report reviewed two patients with advanced melanoma refractory to targeted therapy (BRAF/MEK inhibition) and standard ICI, who subsequently responded to the addition of PARPi to ICI therapy and PARPi to the combination of ICI and BRAFi. PARPi was well-tolerated and demonstrated sustained response for almost a year.

Preclinically, PARPi was found to increase tumor cells’ immunogenicity by creating cytoplasmic chromatin fragments that further activate the cyclic GMP-AMP synthase-stimulator of interferon (cGAS-STING) pathway, increasing cytokine levels, and stimulating interferon signaling in DDR-deficient cells [7]. Additionally, PARPi was implicated in upregulating PD-L1 expression, increasing the tumor mutational burden, altering the tumor microenvironment, and increasing the tumor-infiltrating lymphocytes and genomic instability [8, 9]. These findings could provide a rationale for the patients’ response following the addition of PARPi to ICI therapy. Furthermore, these immune priming effects as well as our clinical findings highlight the potential synergism that may exist between PARPi and ICI therapy.

Clinically, there have been limited reports that highlight the potential benefit in combining ICI with PARPi therapy. A study in an ovarian cancer cohort found that the combination of niraparib and pembrolizumab had resulted in antitumor activity and a similar objective response rate (ORR) regardless of BRCA and HRD status [10]. On contrary, another study that evaluated a triple negative breast cancer cohort found that those with BRCA mutations had a significantly higher clinical response compared to those without BRCA mutations, with an ORR of 47% and a disease control rate (DCR) of 80% vs an ORR of 11% and a DCR of 33%, respectively [11]. However, little clinical data exist in evaluating the combination of PARPi and ICI in melanoma. Importantly, a case report of an advanced melanoma patient previously reported a complete radiologic response on a PET scan of a metastatic liver lesion two months following the administration of nivolumab and olaparib, after disease relapse on maintenance nivolumab [3]. This corresponds well with our findings and highlights the potential clinical benefit that combining PARPi and ICI may bring to advanced melanoma patients with HRD status. Importantly, there is a currently recruiting phase II clinical trial (NCT04633902) evaluating the combination of PARPi and pembrolizumab in patients with advanced melanoma with HR mutations. Results of this trial will especially be important in delineating clinical benefit behind this combinatory approach and highlight any potential drawbacks that it may bring.

Additionally, patient 2 had a BRAF mutation, which prompted the initiation of BRAF/MEK inhibitors in the treatment regimen. The synergism between PARPi and BRAF/MEK inhibitors in melanoma has been previously documented in the preclinical setting. In a study using MAPK inhibitors (MAPKi) resistant melanoma cells, PARPi demonstrated the ability to significantly reduce the migratory and invasive potential of melanoma cells as well as to synergize with MAPK inhibitors through synthetic lethality both in vitro and in vivo [12]. Another study evaluating melanoma cells found that PARPi treatment had restored sensitivity to MAPKi, regardless of HR mutation status, through the reversal of a leading mechanism of resistance known as the epithelial-mesenchymal transition-like phenotype switching [13].

Similarly, Maertens et al. demonstrated that MAPKi had caused a BRCA-ness phenotype in MAPKi sensitive melanoma cells, a known biomarker for PARPi, by suppressing homologous recombination genes. This observation was proven by the cells’ increased response to PARPi as shown by the increased cytotoxicity in MAPKi-sensitive melanoma following PARPi treatment [14]. These preclinical observations provide strong support for the potential synergism that may exist clinically in combining PARPi and BRAF/MAPK inhibitors in melanoma. Clinically, we previously reported three cases of advanced BRAF V600 mutated melanoma patients that exhibited a partial to near-complete response to combinatory PARPi and BRAF/MEK inhibitors following progression on BRAF/MEK inhibitor therapy [15]. These results again correlate well with the response seen in patient 2 and highlight the potential synergism that may exist in combining BRAF/MEK inhibitors and PARPi therapy.

Our case series comes with many limitations, including the inability to discern whether the response seen was due to the synergism of PD-L1 and PARPi therapy, BRAF/MEK inhibitors and PARPi, or PARPi therapy alone in patient 2. Similarly, in patient 1, the response could have incurred because of PARPi alone or the combination of ICI and PARPi. Patient 1 had an ARID2 mutation and patient 2 had a BRCA2 mutation, which are involved in the HR-DDR pathway [16]. Additionally, both patients demonstrated HRD status, which could explain the response seen following PARPi therapy. Furthermore, it is important to note that this is only the case of two patients, which limits the clinical applicability of our results. Nevertheless, this report expands on current literature in highlighting the clinical applicability of PARPi in the setting of advanced melanoma and the potential synergism that may exist in combining the treatment with standard therapies. However, much research is needed to better understand how PARPi could fit in the treatment algorithm against melanoma, as well as to evaluate the potential benefits that may exist in combining this therapy with other treatment options.

Funding Statement

This paper was not funded.

Author contributions

Conceptualization: G.A., G.N. Data Curation: J.P., R.M. Investigation: J.P., R.M. Writing – Original Draft: G.N., R.M., J.P. Writing – Review & Editing: D.C., G.N., G.A., R.M., J.P.

Patient consent

The authors obtained written consent from patients for their photographs and medical information to be published in print and online and with the understanding that this information may be publicly available.

Disclosure statement

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.