Human γδ T lymphocytes are divided in the two main Vδ1^pos and Vδ2^pos subsets on the basis of their TCRδ-chain repertoire. While Vδ1^pos cells preferentially localize in mucosal tissues and skin, Vδ2^pos T cells are mainly enriched in peripheral blood (PB) where they represent about 5% of all circulating T cells. The activation of Vδ2^pos T cells relies on the recognition of non-peptidic compounds (i.e. microbial or stress- or tumor-induced “phosphoantigens”) in association with butyrophilin 3A1 (BTN3A1 also known as CD277). Besides the TCR interactions with phosphoantigens/BTN3A1 complexes, several Natural Killer Receptors (NKRs) are involved in triggering the anti-tumor functions of Vδ2^pos T cells, with the C-lectin type NKG2D playing a major role [1, 2]. The differential expressions of CD27 and CD45RA surface markers identify different Vδ2^pos T cell subsets: CD27^pos/CD45RA^pos naïve cells (T_Naïve), CD27^pos/CD45RA^neg central memory (T_CM), CD27^neg/CD45RA^neg effector-memory (T_EM) and the terminally-differentiated (T_EMRA) CD27^neg/CD45RA^pos cells. These Vδ2^pos T cell subsets diverge not only for their maturation/differentiation status, but also for proliferative capacities, effector functions and resistance to cell death in response to antigens and/or cytokine stimulations [3].

Growing evidences highlighted the high impact of Vδ2^pos T cells in cancer immune-surveillance with promising perspectives in cancer immunotherapy [4, 5]. In this context, two main clinical approaches have been employed to boost anti-tumor activities of Vδ2^pos T cells. The first one activates them through the in vivo administration of either IL-2 or synthetic nitrogen-containing bisphosphonates (NBPs) drugs that, in turn, induce intracellular accumulation of phosphoantigens. A second strategy relies on adoptive transfers of Vδ2^pos T cells expanded in vitro with several methodologies such as the activation with zoledronate [5, 6]. However, these procedures showed both experimental and clinical limits and many efforts are currently being implemented to further improve the effector-functions and persistence in vivo of Vδ2^pos T cells. In this context, cellular senescence is certainly one of the main issues to solve considering that age-related changes of T cells greatly impair their capacity to expand and proliferate, thus leading to dysfunctional immune responses against tumors and pathogens [7]. The shift to senescence and accumulation of mature T cells physiologically occur after 60 years old when both αβ and γδ T lymphocytes lose their co-stimulatory molecules (i.e. CD27 and CD28), acquire terminally-differentiated T_EM and T_EMRA phenotypic profiles, express high constitutive levels of the senescence marker CD57 and shorten their telomerase lengths [8–11]. However, it is still controversial whether CD57 can be used as a single marker to identify senescent Vδ2^pos T cells regardless of differential expression of CD27 and CD45 [3, 11, 12].

Aging is certainly a major burden for social health systems in the industrialized countries as the populations are longer exposed to several pro-tumorigenic risk factors. This leads to a significant higher incidence of cancer onsets in the 6th, 7th and 8th decades of life [13]. Hence, there are rising numbers of elderly patients undergoing anti-cancer conventional chemotherapies (CHT), whose high toxicities greatly hamper both duration and quality of life. In this regard, several lines of clinical and experimental evidence pointed out that these anti-neoplastic treatments further accelerate immune-cell senescence, thus representing negative prognostic factors in aging and worsening the overall clinical outcomes of cancer patients [14, 15].

Since the use of Vδ2^pos T cells is currently considered one of the most promising tools in cancer immunotherapy [4, 5], understanding the exact impact of CHT on their immune-senescence is key to better predict the clinical outcomes of cancer in elderly and to optimize those therapeutic protocols targeting these highly cytotoxic unconventional T cell effectors. Colorectal cancer (CRC) represents the 3rd most frequent solid cancer and more than 50% of CRC patients undergo hepatic dissemination of the primary tumor. The gold-standard therapeutic approach of CRC patients with liver metastasis (CLM) is the surgical removal of hepatic secondary lesions after neoadjuvant combination CHT with or without biological therapy (BT) (Table 1) [16, 17]. Moreover, a higher infiltration of competent immune cells in tumor mass greatly improves the prognosis of CLM patients and increases their overall survival (OS) [18, 19]. Here, we analyze the impact of conventional CHT regimens on the homeostasis and effector-functions of Vδ2^pos T cells in a cohort of CLM elderly patients.Tab1

Neoadjuvant combination chemotherapy (CHT) with or without biological therapy (BT) of enrolled CLM patients

Vδ2^pos T cells were gated within viable CD3^pos/CD45^pos lymphocytes and their absolute counts are significantly lower in the PB of CLM patients underwent CHT compared to those of healthy donors (Fig. 1a-b). We then analyzed the surface expression of CD27 and CD45RA to track the differentiation and distribution of Vδ2^pos T cell subsets. Our data showed a significant increase of Vδ2^pos T_EMRA in CLM patients underwent CHT (28.9 ± 20.6%) compared to healthy controls (9.4 ± 6.4%). This phenomenon is associated with the previous administration of CHT, as the frequency of circulating Vδ2^pos T_EMRA in those CLM patients naïve for CHT (16.7% ±12.6) is similar to that of healthy donors and significantly lower to that of CLM patients underwent CHT (41.6% ±19.6). The increased amounts of Vδ2^pos T_EMRA in CLM patients treated with CHT is counterbalanced by a significant decrease of Vδ2^pos T_CM in the same patients compared to their counterparts naïve for CHT (Fig. 1c-d-e). The great impact of neoadjuvant CHT in shaping the distribution of Vδ2^pos T cell subsets in CLM patients is also confirmed by our findings showing that the number of CHT cycles (8.7 ± 2.7) inversely correlates with the percentages of PB Vδ2^pos T_CM, while not affecting at all the overall frequencies of PB Vδ2^pos T_EMRA (Fig. 1f). This latter dichotomy reflects the different homeostatic status of Vδ2^pos T_CM compared to that of Vδ2^pos T_EMRA, as the first subset is composed of proliferating lymphocytes high susceptible to the toxicity of those chemotherapy compounds that kills all dividing cells without any specificity against tumor blasts. Instead, T_EMRA Vδ2^pos cells are terminally differentiated and not proliferating effectors resistant to CHT, thus explaining their high frequency even after several cycles of neoadjuvant anti-tumor chemotherapies.Fig1Fig. 1

Frequency and distributions of peripheral blood Vδ2^pos T cell subsets in patients affected by liver metastasis of colorectal cancer and underwent chemotherapy. a Representative dot plot flow cytometric graphs showing the gating strategy of viable CD45^pos/CD3^pos/Vδ2^pos T lymphocytes. b Statistical dot plot graph showing the absolute number of CD3^pos (left) and Vδ2^pos (right) T cells per 1 mL of blood in healthy donors (n = 12; mean age: 49.3 ± 9.5) and CLM patients underwent CHT regimens (n = 16; mean age: 51.5 ± 8.1). c-e Representative dot plot graph flow cytometric graph (c) and pie charts (d and e) showing respectively the distribution and the percentages of CD27^pos/CD45RA^pos T_Naive (upper right in dot plot graph and light green in pie charts), CD27^pos/CD45RA^neg central memory (T_CM) (upper left in dot plot graph and gray in pie charts), CD27^neg/CD45RA^neg effector-memory (T_EM) (lower left in dot plot graph and purple in pie charts) and terminally-differentiated CD27^neg/CD45RA^pos (T_EMRA) (lower right in dot plot graph and orange in pie charts) Vδ2^pos T cell subsets. Pie charts compare the frequencies of Vδ2^pos T cell subsets between healthy donors (n = 34; mean age: 51.7 ± 10.8) with age-matched CLM patient underwent CHT (n = 33; mean age: 51.5 ± 8.1) d as well as between CLM patients naïve for CHT (n = 13; mean age: 69.5 ± 8.1) and age-matched CLM patients underwent CHT (n = 41; mean age: 70.1 ± 6.5) (e). f Statistical analysis showing the Pearson correlations between the frequency (%) of either T_CM (left) or T_EMRA (right) Vδ2^pos T cells with the number of CHT cycles (mean number: 8.7 ± 6.5) administered to patients affected by CLM (n = 40)

The present study is aimed to measure the true impact of conventional CHT regimens on unconventional T cell senescence in elderly cancer patients, since the toxicity of conventional anti-tumor therapies greatly impairs their ability to clearance malignant cells [7, 12, 14, 15]. In particular, we focused our investigations on circulating Vδ2^pos cells that are endowed with great anti-tumor potentials currently being targeted by several protocols of cancer immunotherapies [4–6]. Our data showed that CLM patients underwent CHT, although showing lower absolute counts of circulating Vδ2^pos cells, retain high relative frequencies of terminally differentiated and senescent CD57^pos/CD28^neg/CD16^pos T_EMRA Vδ2^pos cells greatly impaired in their effector-functions. This latter subset is resistant to the toxicity exerted by repeated CHT cycles administering DNA-damaging drugs that, in contrast, are highly toxic against less differentiated and still proliferating T_CM Vδ2^pos cells.

The preferential accumulation in PB of senescent CD57^pos T_EMRA Vδ2^pos cells in CLM patients underwent CHT is associated with two major mechanisms. The first one is linked to natural immune-senescence of people aging as the incidence of many cancers is higher in patients ≥ of 60 years old. In this context, liver metastatic CRC is one of the most common causes of cancer deaths worldwide with a higher incidence in elderly [16, 17]. Indeed, our cohort of recruited CLM subjects had a mean age of 61 ± 10.7 years old and both the frequencies of CD57^pos and T_EMRA Vδ2 T cell subsets resulted higher in that fraction of patients older than 60 years. The second mechanism is associated with a direct toxicity of CHT on immune cells, as also highlighted by several studies both in pediatric and geriatric cancer patients [14, 15, 22]. As a matter of fact, we show here that the expression of CD57 on T_EMRA Vδ2^pos cells is much higher on those CLM patients underwent CHT and younger than 60 years old compared to age-matched healthy donors. This demonstrates that neoadjuvant CHT induces immune senescence also on unconventional T cells regardless of CLM patients’ age. Notably, high frequencies of impaired CD57^pos/T_EMRA Vδ2^pos cells were able to persist in PB of CLM patients even after 6 weeks from the completion of the last CHT cycle and before surgical removal of liver metastases. Further prospective studies are required to assess how long senescent and functional impaired Vδ2^pos T cells survive after CHT and what clinical impact they have on the OS of CLM patients. In this regard, it has been already reported that the enrichment of circulating subsets of CD57^pos αβ T cells represents a negative prognostic factor in the clinical outcome of gastrointestinal cancers [23].

Our study also contributes to better characterize immune-senescence of Vδ2^pos T cells, since it has been recently reported that expression of CD57 can define alone their senescent status without the need of also evaluating the expression of both CD27 and CD45RA [11]. This represents a key point that is currently being debated both in physiological and pathological conditions. We found that, at least in a human cancer setting, the expression of CD57 on senescent Vδ2^pos T cell parallels their terminal differentiation towards T_EMRA (CD27^neg/CD45RA^pos), a phenomenon associated with the loss of CD28 and the acquired expression of CD16. These results are in line with a previous study showing that Vδ2^pos T_EMRA are refractory to phosphoantigen stimulation, but rather respond to activation via FcγRIII [21].

The majority of cancer patients are older than 65 years old in line with population aging [14]. In this context, several clinical trials in the elderly are currently being implemented to optimize the anti-tumor activities of unconventional T cells. These therapeutic protocols are mostly aimed to expand Vδ2^pos T cells both in vivo and in vitro [6]. Hence, a better understanding of the mechanisms accelerating immune-senescence in aging is fundamental to boost the effector-functions of these cytotoxic and cytokine-producer T lymphocytes. We show here that, neoadjuvant CHT regimens, although absolutely required to reduce tumor mass in CLM patients before surgery, greatly speed the senescence of Vδ2^pos T cells in synergy with aging of cancer patients. This knowledge will allow us to better optimize immune-therapies against cancers in elderly. Indeed, senescence process can be reversed through the inhibition of p38 mitogen-activated protein kinase (MAPK) signaling [24]. This methodology could be then approached to develop new protocols implementing pre-treatment with MAPK inhibitors in elderly patients with CRC [25]. Alternatively, new methodology can be implemented in vitro to select and expand CD57^neg/Vδ2^pos T cells that better resist to the terminal differentiations and senescence induced by CHT. Further studies are also required to better identify those CHT associated accumulation of impaired and senescent circulating Vδ2^pos T cells.