IL-6 as new prognostic factor in patients with advanced cutaneous squamous cell carcinoma treated with cemiplimab

Cutaneous squamous cell carcinoma (CSCC) is the second most common skin cancer in white-skinned populations. The advanced disease develops in less than 5% of patients, but it is difficult to treat and often has a poor prognosis. Immune checkpoint blockade with cemiplimab, a PD-1 inhibitor, has been approved for advanced/metastatic CSCC [1]. However, criteria for indication of cemiplimab in advanced disease are debated, and predictive factors for initial response to cemiplimab treatment are lacking [2].

CSCC frequently develops in chronically inflamed skin and chronic lesions, including burns, wounds, and ulcers [3]. Soluble inflammatory mediators and growth factors, such as Interleukin 6 (IL-6), IL-8 and tumor necrosis factor-α (TNF-α) are known to promote cell proliferation, inhibit apoptosis, activate survival pathways, and recruit leukocytes to the tumor site [4]. IL-6 is a key inflammatory molecule secreted by M2 macrophages after polarization, mediating the progression of pancreatic cancer [5]. The effect of IL-6 on proliferation, metastasis, and tumor immune evasion of colorectal cancer has been demonstrated in recent years [6]. Based on these considerations, we investigated the prognostic significance of serum IL-6 in patients with advanced CSCC undergoing cemiplimab treatment; serum IL-6 levels assessed at baseline and after treatment with cemiplimab were correlated with oncological outcomes in a retrospective study.

Overall, 39 patients were enrolled. Demographic and clinical characteristics are reported in Table 1. The median age was 76.7 years (range, 39–96 years), and 31 (79.5%) patients were males. Cemiplimab was used as first-line therapy in 28 (71.8%) patients and as second-line in 11 (28.2%) subjects; nine (81.8%) patients had received radiotherapy prior to immunotherapy, one (9.1%) had received targeted therapy, and one (9.1%) had received immunotherapy other than cemiplimab. The most frequent comorbidities were blood hypertension (64%) and angina/coronary artery disease (31%). Eight patients had low serum IL-6 levels, and 31 had high serum levels. At the first follow-up visit, 15 (38.5%) patients achieved PR, 10 (25.6%) reached SD, and 14 (38.5%) were in progression.

No death was observed in the low serum level group, while 14/31 (45.1%) patients with high IL-6 levels died. The mean OS was 17.4 months in the whole population, 20.8 months in the low IL-6 level group, and 16.1 months in the high IL-6 level group (95% CI 13,046–19,184 months). The Kaplan–Meier curve showed a significantly higher OS in the low-level IL-6 group than in the high-level group (Fig. 1).

The mean PFS was 12.4 months in the overall population, 18.9 months in the low-level IL-6 group, and 10.3 months in the high-level group (95% CI 3433–10,133 months) (Fig. 2).

Similar results were obtained in naïve patients treated in first line with cemiplimab. Median OS was higher in the low-level IL-6 group (20.8 months vs 15.6 months95% CI NA), as was PFS (18.9vs 9.6, 95% CI 0.09317–0.7492, p = 0.012) (Fig. 3A, B).

The multivariate analysis showed that serum IL-6 level was correlated with PFS and was an independent predictive factor for response to cemiplimab (Table 2).

The post/pre cemiplimab ratio of serum levels of IL-6 was calculated in 20 patients (Additional file 1: Figure S1). Eight of them had a ratio ≤ 1 and 12 had a ratio > 1 (Additional file 2: Figure S2). The OS was significantly higher in patients with a ratio ≤ 1 (p = 0.028) (Fig. 4). No death was observed in the group with a ratio ≤ 1, and 7/12 (58%) patients died among those with a ratio > 1. The mean PFS was 19.8 months in patients with a ratio ≤ 1, and 10 months in those with a ratio > 1 (95% CI 3800–15,033) (Fig. 5).

Data from this retrospective study showed that high serum levels of IL-6 were associated with poorer survival and shorter PFS in patients with advanced CSCC treated with cemiplimab. In addition, patients whose IL-6 level increased after treatment with cemiplimab, independently of the basal level, had a poorer response to treatment than patients whose level was reduced or stable after immunotherapy.

Our results seem to agree with Li et al. findings showing that plasma levels of IL-6 and IL-8 in cancer patients with metastases were higher than those without metastases (p < 0.05) in a group of 134 patients with different types of cancer. However, according to our knowledge, we are unaware of studies on a cohort of patients with advanced CSCC [9].

On the other end, we are unaware of previous studies demonstrating a predictive role of serum IL-6 in advanced CSCC treated with immunotherapy. Recently, experimental evidence suggested that IL-6 expression may impact the response to immunotherapy. IL-6 is known to be involved in cell proliferation, immune evasion, and progression of squamous cell cancer [6]. Li et al. found that IL-6 deficient (IL-6^−/−) mice were less susceptible to developing experimental colorectal cancer than wild-type mice [6]. Additionally, the blockade of IL-6 activated CD8⁺ T-cell accumulation and led to elevated PD-L1 expression in cancers, suggesting possible sensitization to anti-PD-1 therapy.

IL-6 has been proposed as a possible negative prognostic factor in cancer as it has several effects on immune cell populations which impact tumor development; it regulates myeloid-derived suppressor cells (MDSCs) accumulation and activation, resulting in suppression of anti-tumor T cell and natural killer cell function, and stimulation of tumor cell proliferation, survival, invasiveness, and metastasis [10, 11].

The prognostic and predictive role of IL-6 in patients with metastatic melanoma treated with checkpoint inhibitor immunotherapy has been investigated in three large phase II/III randomized trials [12]. In Checkmate-064 (NCT01783938), in both combined cohorts receiving nivolumab and then ipilimumab or the reverse, baseline IL-6 below the median value (13.3 pg/mL) was associated with better OS compared with above the median. On-treatment levels of IL-6 at week 13 were also associated with OS. Checkmate-066 (NCT01721772) and Checkmate-067 (NCT01844505) showed that high baseline levels of IL-6 were associated with poor prognosis in groups receiving nivolumab, ipilimumab, or dacarbazine. In addition, in all three arms of the Checkmate-067 study, IL-6 was a significant independent variate for OS. Overall, the response to checkpoint inhibitor immunotherapy seems to be related to IL-6 baseline and on-treatment levels in several tumor types, possibly with the activation of common mechanisms (6).

This study has some limitations. The number of patients is limited due to the rarity of the condition and blood samples could not be obtained from some patients during the treatment due to management issues. Additionally, basal and on-treatment serum levels of IL-6 were associated with clinical oncologic outcomes of patients with CSCC treated with cemiplimab, our study was not designed to demonstrate a causal role of IL-6. We nevertheless explored the relationship of IL-6 levels with adverse events (AEs). Patients who had AEs during the treatment had a better OS than patients without AEs (p = 0.04; Additional file 3: Figure S3), while there was not a statistically significant difference for PFS (p = 0.06, Additional file 4: Figure S4). It was further observed that patients with cutaneous AEs (rash or itching) were a subgroup with best outcomes. Additionally, low levels of IL-6 at baseline and on-treatment were associated with the occurrence of cutaneous AEs (baseline n = 13, on-treatment n = 10; p < 0.005 [Mann Whitney]) compared to patients who had not AEs (baseline n = 17, o-treatment n = 9). These data suggest that the activation of non-inflammatory immune mechanisms may be correlated with the development of cutaneous AEs, and more importantly with better treatment response. Among patients with low IL-6 level at baseline, 7/8 were naïve to immunotherapy; this latter feature could stand for a higher probability of response to treatment.

Notwithstanding limitations, as data suggest that IL-6 serum level could be a prognostic factor of the response to cemiplimab, this study opens to new investigations. Further studies should include a larger cohort, and conjointly analyze several inflammatory factors such as IL-8, IL-10 and TNF-α.

In conclusion, we found that changes in serum IL-6 levels after cemiplimab immunotherapy have prognostic significance in patients with advanced CSCC and that high levels at baseline correlate with poor response to treatment.