Notwithstanding the existence of profound immune dysregulation in advanced metastatic breast cancer (BC), it is now well recognized that early disease is also associated with both localized and systemic immune dysfunction (1–4). In the tumor microenvironment (TME), macrophages of the M2-like immunosuppressive phenotype appear to predominate (5–7), where they effectively exclude cytotoxic CD8⁺ T cells from reaching their tumor targets (8). In the case of systemic immunosuppression in early BC, prominent mechanisms include decreased numbers and/or activation of conventional dendritic cells (1, 3, 4), as well as increased numbers of monocytes with an M2-like phenotype (6, 9).

Recently, soluble co-inhibitory immune checkpoint molecules (ICMs) have been implicated as being potential mediators of the systemic immune dysregulation, which is associated with certain types of malignancy (10, 11). Prominent among these soluble co-inhibitory ICMs are: i) cytotoxic T-lymphocyte-associated protein (CTLA-4); ii) programmed cell death protein 1 (PD-1) and its ligand, PD-L1; iii) lymphocyte activation gene 3 (LAG-3); and iv) T cell immunoglobulin and mucin-domain-containing protein 3 (TIM-3) (12). On the other hand, less is known about alterations in the systemic levels of co-stimulatory ICMs measured prior to and following treatment of various types of malignancy. Although the origins of soluble systemic ICMs remain unclear, leakage from the TME seems plausible. In this setting, soluble co-inhibitory ICMs may originate from tumor cells per se, as well as from structural cells such as cancer-associated fibroblasts and cells of the innate and adaptive immune systems recruited to the TME (13, 14). Soluble variant isoforms of the ICMs, which lack transmembrane domains, may arise from alternative RNA splicing, or, alternatively, by proteolytic detachment from cell membranes (10, 12). Importantly, soluble co-inhibitory ICMs appear to retain their biological activities (10, 12).

In the setting of early BC, neoadjuvant chemotherapy (NAC) preceding surgical excision of residual tumor tissue remains a cornerstone of the treatment of this malignancy and strategies to improve the clinical efficacy and safety of this pre-operative procedure are ongoing, including evaluation of the combination of co-inhibitory PD-L1-targeted immunotherapy with chemotherapy (15). Notably, however, the involvement of the various types of soluble ICMs as potential contributors to systemic immune dysregulation in early BC remains largely unexplored, as does the role of NAC and subsequent surgical resection in alleviating cancer-related immune dysfunction.

In accordance with our primary hypothesis, we have therefore compared the plasma concentrations of eight co-stimulatory, six co-inhibitory, and two dual-active ICMs in a cohort of early BC patients with those of a group of healthy control subjects, as well as the effects of NAC and surgery on the plasma levels of these immunoregulatory proteins. This is a follow-up to our previous study in which we reported that the plasma levels of a number of systemic soluble co-stimulatory and co-inhibitory ICMs were significantly decreased in a cohort of early BC patients relative to those of healthy controls, indicative of tumor-related immune dysregulation (16).

Seventy-two female patients of the original cohort (n=98) (16) with early BC attending the Medical Oncology Centre of Rosebank, Johannesburg, South Africa, who had completed treatment (NAC) and surgery were deemed eligible for recruitment to the study. Eligibility criteria included: i) age ≥18 years; ii) Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0, 1, or 2; iii) histologically confirmed BC (17) classified as clinical stage I, II or III according to the AJCC Breast Staging 8th Edition (18); iv) normal bone marrow, liver and renal function; v) NAC with anthracycline (A) and/or a taxane (T)-based chemotherapy regimen. Human epidermal growth factor receptor 2 (Her2) -positive patients received neoadjuvant trastuzumab-based treatment.

Exclusion criteria were as follows: i) prior systemic chemotherapy for BC within five years; ii) a history of any other malignancy during the preceding five years, with the exception of basal cell or squamous cell carcinoma of the skin treated with local resection only, as well as carcinoma in situ of the cervix; iii) patients with confirmed stage IV disease; iv) known seropositivity for human immunodeficiency virus (HIV) and/or hepatitis B or hepatitis C viruses; v) uncontrolled intercurrent illness including, but not limited to, active infection, symptomatic congestive heart failure, unstable angina pectoris, cardiac arrhythmia or psychiatric illness/social situations that would limit compliance with study requirements; vi) pregnancy or breast feeding; and vii) otherwise not deemed to be a good study candidate according to the sole discretion of the principal investigator (BLR).

The group of healthy female control participants (n=45, mean age = 49.9 years; range 24–70 years) was recruited almost exclusively from the female personnel of the Medical Oncology Centre of Rosebank and the Faculty of Health Sciences, University of Pretoria. Exclusion criteria included: i) those with uncontrolled medical conditions; and ii) any potential participant deemed unwell by the qualified nursing sister in attendance on the day of venepuncture. Immunohistochemical staining was performed for ER, PR, Her2, and Ki67. Fluorescence in situ hybridization (FISH) was used to confirm Her2 positivity. Clinical assessment of the primary tumor and lymph nodes was done using bi-dimensional caliper measurements of the primary tumor and axillary nodes. The lymph node positivity at presentation was assessed clinically, radiologically, or by sentinel lymph node biopsy. Sonographic assessments of the primary tumor and lymph nodes were performed regularly. The pathological complete response (pCR) was defined as the complete disappearance of invasive cancer in the breast and the absence of tumor cells in the axillary lymph nodes.

Permission to undertake this study and to draw blood from patients with early BC and from matched, healthy control subjects was granted by the Research Ethics Committee of the Faculty of Health Sciences, University of Pretoria, in full compliance with the World Medical Association Declaration of Helsinki 2013. Two submissions were approved: firstly, the BC study, and secondly, a submission in respect of the healthy control subjects (respective Approval Numbers 517/2017 and 762/2020). Prior, written informed consent was obtained from all participants.

Whole blood samples of 20 mL and 10 mL volumes were collected in ethylenediaminetetraacetic acid (EDTA)-containing vacutainers from the BC patients and control subjects, respectively, and the plasma promptly separated, aliquoted and stored at -80°C until the plasma concentrations of the various soluble co-inhibitory and co-stimulatory immune checkpoints were measured.

In a forerunner to the current study, we recently reported that the plasma concentrations of a large group of soluble ICMs, comprising both co-stimulatory and co-inhibitory proteins, were significantly decreased in treatment-naïve, newly-diagnosed patients with early BC (n=98), relative to those of a cohort of healthy control subjects (n=45) (16). The current study is a follow-up to our earlier investigation and was designed to explore the immune-restorative potential of implementation of NAC and subsequent surgical resection of residual tumor tissue in the same, albeit somewhat smaller, cohort of early BC patients who completed both NAC and surgery (n=72). In terms of realization of our primary objectives, the following represent the most significant and novel findings: i) the plasma concentrations of five soluble co-stimulatory ICMs (CD27, CD40, GITRL, ICOS, GITR) were significantly lower than those of the control group pre-treatment, while those of CD28, CD80 and CD86 were numerically decreased, but not significantly so; ii) post-NAC, the plasma concentrations of six soluble co-stimulatory immune checkpoints (CD27, CD28, CD40, CD80, ICOS, GITR), all of which are involved in the activation of anti-tumor cytotoxic T cells (12), normalized; iii) recovery of the plasma levels of these six ICMs remained essentially unchanged following surgical resection; iv) the levels of three soluble co-inhibitory ICMs (PD-L1, LAG-3 and TIM-3) were also significantly lower in the cohort of early BC patients and increased significantly in magnitude relative to pre-treatment levels post-NAC, with no further augmentation post-surgery. Notably, and somewhat concerningly, the plasma levels of these three co-inhibitory ICMs were significantly higher, especially in the case of LAG-3 and TIM-3, than those of the control group.

Of the other soluble co-inhibitory ICMs, BTLA and CTLA-4 did not respond to NAC, remaining significantly decreased relative to their pre-treatment values, while the median level of soluble PD-1, which was comparable in both groups, remained unchanged. In the case of the persistently low levels of soluble CTLA-4, it is noteworthy that taxane has been reported to deplete the numbers of regulatory T cells (Tregs), a major cellular source of this ICM (19). However, further follow-up studies are necessary to establish if soluble CTLA-4 levels increase over time. Of the two dual-active soluble ICMs, the pre-treatment level of HVEM was significantly lower than that of the corresponding median control value, recovering post-NAC to values, which were considerably higher than those of the control group. In the case of soluble TLR-2, the median pre-treatment value for this dual-active ICM was comparable with that of the control group, increasing modestly, albeit significantly, post-NAC.

With respect to potential mechanisms that contribute to the increased systemic concentrations of the six soluble co-stimulatory ICMs following implementation of NAC in our cohort of early BC patients, it is noteworthy that the components of NAC, namely anthracycline, taxane, cyclophosphamide and platinum-based agents, are all effective inducers of immunogenic cell death, a process which restores anti-tumor immunity in the TME (20, 21). In this context, our findings are somewhat similar to a recently published study by Dong et al. (22). These authors reported that administration of the immunomodulatory agent, sorafenib, to patients with advanced hepatocellular carcinoma (n=53) resulted in significant increases in the median plasma concentrations of 11/16 of the soluble co-inhibitory and co-stimulatory ICMs tested, an effect which was evident after two weeks of therapy. Differences between our study and that of Dong et al., include the types of malignancy investigated, as well as types and duration of anticancer therapy; another notable difference between the two studies relates to the apparent absence of a group of healthy control subjects in the study reported by Dong et al. (22).

Given that the increases in the median plasma concentrations of the soluble co-stimulatory ICMs, together with associated clinical improvement in our BC patient cohort, may be linked to immunogenic cell death induced by NAC, the corresponding increases in the levels of the soluble co-inhibitory ICMs, PD-L1, LAG-3 and TIM-3, relative to those of the control group, do seem somewhat counterintuitive. Although this may relate to restoration of a balance in immune homeostasis, the unexpectedly high increases in the plasma levels of these co-inhibitory ICMs may, however, compromise the efficacy of therapeutic monoclonal antibodies (mAbs). In this context, it is noteworthy that persistently elevated plasma levels of soluble LAG-3 and TIM-3, like PD-L1, are associated with decreased survival in patients with various types of advanced malignancies (23–26).

Current approaches in the treatment of early BC include NAC, which in the case of TNBC in particular, may also incorporate PD-1/PD-L1 blockading mAbs as mentioned above (15). Against this background, two recently published, randomized, placebo-controlled clinical trials in early TNBC are noteworthy, namely the IMpassion031(n=333 patients) and KEYNOTE-522 (n=1174 patients) trials, which evaluated the clinical efficacy of addition of co-inhibitory ICM-targeted immunotherapy with either atezolizumab (PD-L1-targeted mAb) or pembrolizumab (PD-1-targeted mAb) in combination with NAC, relative to that of placebo + NAC, respectively (27, 28). The authors reported significantly improved pCR responses and event-free survival rates, irrespective of tumor PD-L1 positivity, in the Impassion031 and KEYNOTE-522 clinical trials, respectively (27, 28). These observations have been confirmed in two recent systematic reviews and meta-analyses. The first of these, which encompassed six randomized controlled trials focused on early TNBC (n=2142 patients), reported that NAC in combination with PD-1/PD-L1 inhibitors also resulted in improved pCR rates and event-free survival (29). The second included eight randomized controlled trials, encompassing 4901 patients with both early and metastatic TNBC (four trials in each category) (30). The authors of this study reported improved OS and PFS in patients treated with immunotherapy together with chemotherapy relative to those treated with chemotherapy alone (30). Taken together, these various studies are consistent with a beneficial, immune-restorative interaction between NAC and co-inhibitory ICM-targeted immunotherapy.

Although these and other studies document significant advances that have been made in treating TNBC patients (27–37), the magnitude of this benefit, as well as long-term survival, do, however, remain relatively low. Moreover, most of the clinical research in the field of immunotherapy has been focused on targeting PD-L1. Limitations of our study include sample size, diversity of BC subtypes and data collected from a single institution. Moreover, different therapeutic PD-1/PD-L1 therapeutic monoclonal antibodies require different companion tests. Some assays involve determination of PD-L1 levels on the tumor cells, some on immune cells and some on both cell types. An additional caveat is that only PD-L1 is extensively used in clinical practice. Problematically, however, some tumor types respond to immunotherapy independently of PD-L1 expression. Clearly, additional research of the type described in the current study, investigating additional checkpoints and additional pathways is required to improve long-term outcomes. These include discerning insights into the mechanisms of resistance against immune checkpoint blockade and the cell types involved, as well as the identification of novel, reliable targets and biomarkers to predict responsiveness to treatment. Importantly, the results of the present study indicate, that co-inhibitory ICMs other than PD-L1, specifically LAG-3 and TIM-3, are also prominent in early BC patients following implementation of NAC and should therefore be considered as potential targets for future investigation in these patients as an adjunctive strategy to improve outcomes.

We previously described a state of systemic immune quiescence and immune dysregulation in patients with early BC, which was associated with decreased levels of soluble co-stimulatory and co-inhibitory ICMs. In the current follow-up study, we focused on the effect of NAC and surgery on the levels of these various ICMs. Novel findings originating from the current study include substantial increases in the levels of the majority of the soluble ICMs post-NAC, which did not change significantly post-surgery. Improvement in the levels of six of the co-stimulatory ICMs is indicative of immune- restoration, as all of these are involved in the activation of anti-tumor cytotoxic T cells. In the case of the co-inhibitory ICMs, we observed substantial increases in the levels of LAG-3, TIM-3 and PD-L1, identifying these ICMs as potential future targets in the management of TNBC. Although none of the 16 measured soluble plasma ICMs correlated with the pCR, an extended follow-up study involving a larger number of early BC patients is necessary to demonstrate possible associations of these soluble ICMs with OS and PFS.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Permission to undertake this study and to draw blood from patients with early BC and from matched, healthy control subjects was granted by the Research Ethics Committee of the Faculty of Health Sciences, University of Pretoria, in full compliance with the World Medical Association Declaration of Helsinki 2013. Two submissions were approved; firstly, the BC study, and secondly, a submission in respect of the healthy control subjects (respective Approval Numbers 517/2017 and 762/2020). Prior, written informed consent was obtained from all participants. The patients/participants provided their written informed consent to participate in this study.

BR, HS, SN, and RA conceptualized the study. BR, CB, and SN provided clinical oversight. HS, AT, NH, TS, LK, and LH were responsible for diligent processing and storage of blood/plasma specimens and for performance of laboratory analyses of plasma specimens. BR, PM, and TS advised on aspects of statistical analysis of data. All authors contributed to the article and approved the submitted version.