Enormous scientific progress in molecular tumor biology and immune surveillance has resulted in the introduction of effective targeted therapies [e.g., by blocking the hyperactive B-rapidly accelerated fibrosarcoma (BRAF)‐mitogen-activated protein kinase (MEK) pathway that represents a hallmark of melanoma] and immunotherapies [including single or dual blockade of cytotoxic T lymphocyte-associated protein-4 (CTLA‐4) and programmed cell death protein-1 (PD-1)] for metastasized melanoma (1–4). An increasing number of clinical trials have impressively shown that BRAF inhibitors (BRAFi; such as dabrafenib, vemurafenib, and encorafenib), MEK inhibitors (MEKi; such as trametinib, cobimetinib, and binimetinib), and immune checkpoint inhibitors [ICIs; including CTLA-4 inhibitors (CTLA-4i) and PD-1 inhibitors (PD-1i), such as ipilimumab and pembrolizumab, nivolumab, respectively] improve clinical outcomes when compared to chemotherapy (1, 3). The relevance of the immune system in controlling melanoma is generally accepted, and since the introduction of CTLA-4 single blockage (ipilimumab) in 2011, clinical application of ICIs for melanoma treatment has shown for the first time in history the impressive therapeutic potential of drugs that regulate immunosurveillance (1, 3). Systemic treatment of advanced melanoma has moved forward rapidly, and although these new treatment modalities have completely changed the reality from one of poor responses and short survival to a very promising new clinical picture of impressively high response rates, greatly prolonged disease control, and the possibility of aiming at the ultimate goal of a cure for some patients (1, 3), treatment of advanced melanoma is still challenging.

Consequently, there is urgent need to characterize the molecular mechanisms that exert the therapeutic efficacy of these new treatment modalities and to identify factors that show promise to further optimize melanoma treatment. It has been shown that low levels of vitamin D are associated with poorer prognosis, thicker tumors, ulceration, and increased inflammation in primary melanomas (5, 6). Several independent lines of investigation have supported the hypothesis that 25(OH)D serum concentration (s.c.) may be associated with the risk and prognosis of melanoma. We reported in 2009 in a pilot study (7) significantly reduced 25(OH)D s.c. in stage IV (median 13.10 ng/ml, n = 115) as compared to stage I (median 16.40 ng/ml, n = 50) melanoma patients (p = 0.006). In that investigation, a trend toward a greater tumor thickness of the primary cutaneous melanomas was seen in patients with 25(OH)D s.c. <10 ng/ml (median: 2.55 mm) as compared to >20 ng/ml (median: 1.5 mm), although this difference was not statistically significant (p = 0.078). The patients with low 25(OH)D s.c. (<10 ng/ml) had earlier distant metastatic disease (median: 24.37 months) as compared to those with 25(OH)D s.c. >20 ng/ml (median: 29.47 months), although this difference was also not statistically significant (p = 0.641) (7). Newton-Bishop et al. reported in the same year a significant association between higher vitamin D level and lower Breslow primary tumor thickness as well as lower rates of relapse and death in a two-stage study (8). In 2014, we published an affiliating investigation (9), demonstrating significantly (p  =  0.004) lower 25(OH)D s.c. in melanoma patients (median  =  13.6 ng/ml) as compared to controls (median   = 15.6 ng/ml). Primary tumors of patients with low 25(OH)D s.c. (<10 ng/ml) had significantly (p  =  0.006) greater Breslow thickness (median: 1.9 mm) as compared to patients with higher levels (>20 ng/ml; median: 1.00 mm). In that study, patients with 25(OH)D s.c. in the lowest quartile had inferior overall survival (OS; median: 80 months) compared with the highest quartile (median: 195 months; p  =  0.049).

The experimental support for the concept that vitamin D status is of relevance for progression of melanoma and other malignancies was the demonstration of the almost ubiquitous expression (including in melanocytes, melanoma cells, and many immune cells) of both the vitamin D receptor (VDR) and the activating enzyme (vitamin D-1α-hydroxylase, CYP27B1), which converts serum 25-hydroxyvitamin D [25(OH)D] into the biologically active vitamin D metabolite 1,25-dihydroxyvitamin D [1,25(OH)₂D] (10, 11).

Vitamin D compounds exert many potent antitumor effects on melanoma and other malignancies that involve regulation of proliferation, differentiation, apoptosis, angiogenesis, stemness of cancer stem cells, tumor microenvironment, and various immune functions (12) via a complex network of cellular signaling pathways. Binding and transcriptional activation of VDR by 1,25(OH)₂D regulate multiple pathways [including VDR-, peroxisome proliferator-activated receptor (PPAR)-, p53-, Wnt/β- (The term Wnt results from the fusion of the name of the Drosophila segment polarity gene termed wingless and the name of the vertebrate homolog, integrated or int-1), catenin-, and mitogen-activated protein kinase (MAPK)- mediated signaling] and functions that are implicated in carcinogenesis, cancer progression, metastatic potential, and immunosurveillance (12). Vitamin D-mediated immunomodulating activities 10,12] that have been widely explored in autoimmune disorders include inhibition of T-helper lymphocytes [often involved in the pathogenesis of immune-related adverse events (iAEs) that complicate ICI treatment], upregulation of Programmed death-ligand-1 (PDL-1) expression on both epithelial and immune cells (suggesting a synergic therapeutic effect in combination with ICIs, for which further investigation is needed), and stimulation of antimicrobial peptides in macrophages and other immune cells (increasing the efficacy of antibody-based cancer therapy) (9, 13).

In summary, an increasing body of evidence now convincingly indicates that 25(OH)D s.c. is of high importance for the clinical outcome in melanoma patients. However, there are limited data of the impact of 25(OH)D s.c. on clinical outcome in advanced melanoma. Considering the growing interest in comparative effectiveness research (CER) and to gain further insights into the relevance of serum 25(OH)D concentrations on prognosis of melanoma, we have now carried out this side-by-side analysis in a subgroup of patients that is at present of particular importance, namely, patients treated for metastasized disease with BRAFi/MEKi (monotherapy with BRAFi—dabrafenib, vemurafenib, encorafenib—and/or combination therapy with BRAFi and MEKi—dabrafenib/trametinib, vemurafenib/cobimetinib, and encorafenib/binimetinib) and/or immunotherapy (mono- and or combination therapy with CTLA-4i—ipilimumab—and/or PD-1i—pembrolizumab, nivolumab).

We show in this comparative, non-interventional, side-by-side analysis that low vitamin D status is a strong predictor of poor outcome in patients treated for advanced melanoma with BRAFi, MEKi, CTLA-4i, and/or PD-1i. Investigating both “baseline” and “average” 25(OH)D s.c., OS was markedly reduced in severely vitamin D-deficient individuals as compared to patients with 25(OH)D s.c. ≥10 ng/ml.

Results of our study are well in line with other investigations that support the concept that vitamin D deficiency may favor poor outcome in cancer patients.

Recently, it has been shown that vitamin D regulates via different pathways and independent mechanism functions of many cell types involved in immunologic reactions and in the antitumor response (17–27). It was demonstrated that vitamin D is not exclusively activated through consecutive hydroxylations by cytochrome P450 family enzymes at positions C25 (by CYP2R1 or CYP27A1) and C1 (by CYP27B1) to produce 1,25(OH)₂D₃ but also through the action of CYP11A1 to produce mono-, di-, and trihydroxy-D₃ products that can be further modified by CYP27B1, CYP27A1, and CYP24A1 (17–27). It has been shown that these alternative hydroxyderivatives are circulating in serum (26, 27). The active forms of vitamin D₃, in addition to regulating calcium metabolism, exert pleiotropic activities, which include anticarcinogenic and anti-melanoma effects in experimental models, with photoprotection against UVB-induced damage (17–27). These diverse effects are mediated through an interaction with the VDR and/or as most recently demonstrated through action on retinoic acid orphan receptors (ROR)α and RORγ (21). With respect to melanoma, low levels of 25(OH)D are associated with thicker tumors and reduced patient survival (7, 8, 17, 18). Furthermore, single-nucleotide polymorphisms (SNPs) of VDR and the vitamin D-binding protein (VDP) genes affect melanomagenesis or disease outcome (17–27). Clinicopathological analyses have shown positive correlation between low or undetectable expression of VDR (20, 21) and/or CYP27B1 (22) in melanoma with tumor progression and shorter OS and disease-free survival (DFS). Paradoxically, this correlation was reversed for CYP24A1 (inactivating 24-hydroxylase) (18), indicating that this enzyme, while inactivating 1,25(OH)₂D₃, can activate other forms of vitamin D₃ that are products of the non-canonical pathway initiated by CYP11A1 (25). An inverse correlation has been found between the levels of RORα and RORγ expression and melanoma progression and disease outcome (21). It has been proposed that defects in vitamin D signaling including vitamin D₃ activation/inactivation, and the expression and activity of the corresponding receptors, affect melanoma progression and the outcome of the disease (17). Consequently, it has been postulated that the existence of multiple bioactive forms of vitamin D₃ and alternative receptors affecting the behavior of melanoma should be taken into consideration when applying vitamin D management for melanoma therapy. In this context, it would also be interesting to know whether expression levels of VDR and/or of other proteins involved in mediating vitamin D-induced cellular signaling, or SNPs in their corresponding genes, may have an impact on the clinical outcome in patients with advanced melanoma.

Vitamin D deficiency-induced alterations in immune function, that may be of relevance for melanoma progression, include the upregulation of PDL-1 expression on both epithelial and immune cells. This finding indicates a synergic therapeutic effect in combination with ICIs, that is well in line with results of this study but needs to be further investigated. Vitamin D deficiency-induced alterations in immune function include impaired function of tumor-associated macrophages (TAMs) that may result in reduced efficacy of antitumor therapy. Infiltration by macrophages represents a characteristic morphological hallmark in many malignancies. In theory, macrophages and other immune cells involved in immune surveillance of solid tumors are able to kill malignant cells. However, melanoma cells (and many other cancer cells) are capable of escaping from immunosurveillance, and for reasons that are not yet well understood, TAMs in general do not kill them. Although macrophages can, in principle, target neoplastic cells and mediate antibody-dependent cellular cytotoxicity (ADCC), TAMs regularly fail to exert direct cytotoxic functions. Until today, the underlying mechanisms responsible for this observation remain unclear. Interestingly, Bruns et al. (9) demonstrated that inflammatory M1 macrophages kill proliferating high-grade B-cell lymphoma cells by releasing the antimicrobial peptide cathelicidin in a vitamin D-dependent fashion. They further showed that cathelicidin directly induces cell death by targeting mitochondria of these malignant cells. In contrast, anti-inflammatory M2 macrophages and M2-like TAMs exhibit an altered vitamin D metabolism, resulting in a reduced production of cathelicidin and consequently in inability to lyse malignant cells. In that study, treatment of M2 macrophages with the bioactive form of vitamin D, 1,25(OH)₂D₃, or a VDR agonist effectively induced cathelicidin production and triggered tumoricidal activity. Furthermore, rituximab-mediated cytotoxicity of vitamin D-treated M2 macrophages was cathelicidin-dependent. Finally, vitamin D treatment of 25(OH)D-deficient volunteers in vivo or primary TAMs in vitro improved in that study rituximab-mediated ADCC against B-cell lymphoma cells. These data indicate that activation of the vitamin D signaling pathway stimulates the antitumor activity of TAMs and improves the efficacy of ADCC. In summary, Bruns et al. (9) showed that vitamin D can help promote the antitumor activity of macrophages and stimulate their production of cathelicidin, an antimicrobial peptide that can also induce tumor cell death. The results suggest that for melanoma and other cancer patients who are deficient in vitamin D, providing vitamin D supplementation may be helpful in battling the disease and promoting the efficacy of antitumor therapy. It can be speculated that the poor outcome that we here report in patients treated for advanced melanoma with combination therapy with BRAFi and MEKi and/or with CTLA-4i or PD-1i immunotherapy may be at least in part mediated by vitamin D deficiency-induced impairment of immune function.

However, it has to be noted that the associations that we show in this observational study are no definite proof of a causal relationship. Dissociating the effects of overall fitness on responses to treatment vs. the effects of vitamin D status per se is difficult. In this context, we cannot exclude reverse causality, indicating that 25(OH)D s.c. at least in part might be a consequence of disease progression rather than causal.

It cannot be excluded that the association of vitamin D status with clinical outcome that we report in this study may at least in part be caused by other factors (e.g., “good health,” resulting in frequent outdoor activities) for which vitamin D status may serve as a marker for. It is well known that vitamin D status is positively associated with factors may have in turn a potential effect on melanoma survival. However, it has to be noted that when potentially interfering factors, including location of metastases (e.g., in liver or kidneys, representing organs of high importance for the vitamin D endocrine system), sex, age, history of other malignancies, diabetes mellitus, hyperlipidemia, hypercholesterinemia, coronary heart disease, heart insufficiency, heart rhythm disorders, hypertension, kidney, thyroid or liver disease, asthma bronchiale, ethanol and nicotine abuse, were assessed to adjust Cox regression modeling, none of them were identified as confounders ( Supplementary Table S1 ).

Other limitations of our study include the small study size. Although case numbers in some of our analyses, including subgroup analyses, were low, subgroup analyses and results after adjusting for various parameters, including BRAF/N-RASmut (HR: 0.138, p = 0.000004) and markers of TL such as average LDH (HR 0.306, p = 0.026) and S100P (HR 0.225, p = 0.003), indicate that the relationship between vitamin D status and clinical outcome in melanoma patients treated for metastasized disease with BRAFi/MEKi and/or immunotherapy may at least in part be causal. However, one has to keep in mind that the small sample size of our observational study makes it difficult to draw definite conclusions and also makes it difficult to interpret the results of subgroup analyses, e.g., looking at BRAFmut vs. BRAFwt melanomas.

Driver mutations in BRAF are found in approximately 50% of melanomas, play an important role in cell proliferation and metastases, and also increase inflammation by promoting secretion of pro-inflammatory cytokines, including interleukin (IL)‐6 and IL‐8 (5). Inflammation aids proliferation, survival of cancer cells, angiogenesis, and metastasis. However, there are limited data that analyze the relationship between BRAF and vitamin D status in patients with melanoma.

A retrospective study determined levels of serum 25(OH)D in patients with primary and metastatic melanoma and made comparisons according to age, sex, stage, season, and BRAF status (5). In that study, BRAF status was established in 93 patients, 59 of whom (63%) were positive for BRAF mutation. Of these, 22 patients (37%) were severely vitamin D-deficient compared with three (9%) BRAF wild‐type (P = 0.006). Of the 43 patients who died, 37 (86%) were vitamin D-deficient; 14 of the 43 deceased patients (33%) were severely vitamin D-deficient and 11 of these 14 patients (79%) were positive for BRAF mutation.

Studies have demonstrated a 3-fold increased risk of death for higher‐risk tumors harboring NRAS or BRAF mutations compared with wild‐type melanomas, after adjusting for other prognostic factors (28). Further studies are required to investigate the association of severe vitamin D deficiency with BRAF‐mutated melanoma, especially as low vitamin D levels may impact and increase the incidence of immune‐related adverse events in patients receiving immunotherapy (29). In this study, over 80% of patients with melanoma were vitamin D deficient; therefore, we recommend testing for all patients with melanoma regardless of stage.

Here we report that low vitamin D status is a strong predictor of poor outcome in patients treated for advanced melanoma with BRAFi, MEKi, CTLA-4i, or PD-1i, whether or not the tumors harbor mutations in the BRAF gene. However, vitamin D status-induced effects on outcome appear to be pronounced in individuals with BRAF mutant melanomas. In the subgroup of patients whose melanomas do or do not harbor mutations in the BRAF gene, individuals with average 25(OH)D serum concentrations ≥10 ng/ml had over the complete OP a 72.5% (p = 0.63) and 85.2% (p = 0.000134), respectively, reduced risk to die, as compared to individuals with average 25(OH)D serum concentrations <10 ng/ml.

Immunomodulating activities of vitamin D compounds include inhibition of T-helper lymphocytes that are often involved in the pathogenesis of immune-related adverse events (iAEs) that complicate ICI treatment, upregulation of PDL-1 expression on both epithelial and immune cells, suggesting a synergic effect in combination with ICIs, for which further investigation is needed.

Although the results were not statistically significant, our study indicates that the vitamin D status-associated improved outcome in our patients may be associated with a decreased frequency of AEs. During the complete OP, a trend for slightly reduced AEs (AE ratio -1.01) was documented in patients with average 25(OH)D serum concentrations ≥10 ng/ml as compared to severely vitamin D-deficient individuals [average 25(OH)D serum concentration <10 ng/ml] (p = 0.449). If this trend will be confirmed in future investigations, it can be speculated whether this decrease in the frequency of AEs may be caused by effects of vitamin D compounds on melanoma cells, immune cells, or on other cell types.

Interestingly, results of our study indicate that patients with BRAFmut melanomas experience more AEs than patients with BRAFwt melanomas. In our study, patients with BRAFmut melanomas and with average 25(OH)D s.c. ≥10 ng/ml showed a trend for a higher frequency of AEs, (0.326, p=0.289) as compared to severely vitamin D-deficient individuals [mean 25(OH)D s.c. <10 ng/ml]. In patients with BRAFmut melanomas, an increase in average 25(OH)D s.c. of 1 ng/ml was associated with an increase (trend) of the frequency of AEs (AE ratio +0.432, 0.001, p=0.765 p = 0.266). In contrast, patients with average 25(OH)D s.c. ≥10 ng/ml and with BRAFwt melanomas showed a trend toward a lower frequency of AEs as compared to severely vitamin D- deficient individuals. In patients with BRAF wt melanomas, an increase in average 25(OH)D s.c. of 1 0.001, p=0,765 ng/ml was associated with a non-significant decrease of the AE-ratio (-0.012, p = 0.137). It can be speculated whether these findings may—directly or indirectly—be caused by BRAF-induced effects on cancer and immune cells that are of importance for the therapeutic efficacy in melanoma.

These immune-modulating activities include inhibition of T-helper lymphocytes, notoriously involved in the pathogenesis of immune-related AEs (iAEs) that complicate ICI treatment, upregulation of PDL-1 expression on both epithelial and immune cells, suggesting a synergic effect in combination with ICIs, for which further investigation is needed. In this context, other factors including survivor bias that is caused by the fact that the longer people survive, the longer they have time to develop toxicities also need to be considered.

In the UK and some other countries, analysis of serum 25(OH)D concentration is recommended by the National Institute of Clinical Excellence for any new patient diagnosed with melanoma (30) but not for those with already established or metastatic disease. Based on the data presented, we would suggest to measure Vitamin D and at least avoid clear deficiency by appropriate substitution.

Further studies are warranted to elucidate the intricacies of a potential direct molecular interference of the vitamin D pathway and BRAF- or ICI-modulated pathways and intercellular signaling. A causal role of vitamin D in better outcome will only be answered by interventional studies like our group has previously pioneered in the treatment of high-grade lymphomas (29).

To take part in the Resource Identification Initiative, we used the corresponding catalog number and RRID in our article.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Ethik-Kommission der Ärztekammer des Saarlandes Faktoreistr.4 66111 Saarbrücken, Germany. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

Substantial contributions to conception and design: JR, FB, and JS. Acquisition of data and/or analysis and interpretation of results: JR, FB, JS, SW, RS, and CP. Substantial scientific and intellectual contributions to the drafting or rewriting of the initial and/or revised article: JR, FB, JS, SW, TV, RS, and CP. Approval of the final accepted version of the article: JR, FB, SW, JS, CP, RS, and TV.

Saarland University, with JR and TV as principal investigators, received funding from the Jörg Wolff Foundation (Stuttgart, Germany). JR received speaker’s honoraria from Leo Pharma (Neu-Isenburg, Germany) and Cogitando (Neunkirchen, Germany).

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.