Liquid Biopsy: a new tool to overcome CDKi resistance mecha- nism in Luminal metastatic breast cancer

Breast cancer is the most commonly diagnosed cancer in women worldwide. Approximately, 70 % of breast cancer patients express hormone receptors (HR) (Luminal subtype). Adjuvant endocrine treatments are the standard of care in HR+/HER2breast cancer. Over time, about 50% of those patients develop endocrine resistance and metastatic breast cancer. Cyclin-dependent kinase inhibitors (CDKi) in combination with an aromatase inhibitor or fulvestrant have demonstrated superior efficacy increasing progression-free survival, with a safe toxicity profile, in HR+/HER2metastatic breast cancer patients. CDKi blocks kinases 4/6 ATP-binding domain preventing G1/S cell cycle transition. Despite this, not all patients respond to CDKi and those who respond, finally develop resistance to combination therapy. Different studies, in tumour tissue or cell lines, have tried to elucidate the mechanisms underlying this progression, but there are still no conclusive data. In the last few years, liquid biopsy has contributed relevant information to this knowledge. Liquid biopsy can be performed in real-time, non-invasively and be repeated whenever needed. Circulating tumour material are potential prognostic markers in metastatic luminal breast cancer to determine patient prognosis, monitor disease and treatment selection. The objective of this review is to outline the different studies carried out in HR+ metastatic breast cancer patients treated with CDKi plus endocrine therapy using liquid biopsy approaches looking for possible resistance mechanisms.


Introduction
Breast cancer (BC) is the most common cancer and the second cause of cancer death among women worldwide, mainly to metastasis. The incidence is more than 2 million people with a 5-year prevalence [1]. There are different breast cancer subtypes: hormone receptor (HR+) (luminal A and luminal B), human epidermal receptor 2 (HER2) enriched, and basal-like or triple-negative (TNBC), defined by the lack of estrogen (ER) and progesterone receptor (PR) and the absence of HER2 protein overexpression [2]. The Luminal subtype represents approximately 70% of the cases and it is characterized by the expression of hormone receptors; estrogen and progesterone [3]. Hormone receptor-positive (HR+) patients are treated with endocrine therapies (ET) such as tamoxifen, anastrozole, letrozole or exemestane, to block the hormone receptor or inhibit systemic estrogen production. Most of the patients diagnosed with primary luminal BC are treated by ET adjuvantly to surgery and or radiotherapy [4], but it was estimated that 30 % of patients will develop metastasis, while 6% already have metastatic disease at the time of diagnosis [5], [6].
Different mechanisms of endocrine resistance have been identified such as upregulation of cyclins, cyclin-dependent kinases (CDKs) and mitogen signalling pathways (PI3K and RAS), reduction of CDK inhibitory proteins (p16, p21, p27), mutations or loss of ESR1 as well as epigenetic alterations [7]. CDKs act downstream of estrogen signalling, controlling cell cycle progression. As these proteins are normally altered in breast cancer, have been considered a key target for therapeutic intervention in the metastatic setting [6].
The Food and Drug Administration (FDA) and The European Medicines Agency (EMA) approved the combination of cyclin-dependent kinase inhibitors (CDKi) with endocrine therapy to treat advanced luminal BC. So far three different inhibited cyclins called palbociclib, ribociclib and abemaciclib have been marketed. Clinical trials [8][9][10][11][12][13] have demonstrated that the CDK4/6 inhibitors plus endocrine therapy (ET) combination improves progression free survival and survival rates compared with ET alone. Nevertheless, not all patients respond to CDKi and those who initially respond, ultimately progress. There are factors responsible for endocrine resistance that have not yet been identified [14], which complicates the study of resistance combining both therapies. In addition, preclinical studies suggest that the heterogeneous resistance to CDKi further hinder decipher resistance mechanisms for the combinatorial therapy [15]. The efforts made using mainly primary tumour tissue samples or cell lines did not get conclusive results. Therefore, a change in the paradigm is needed for this emergent drug-resistant patients group. Precision oncology throughout the analysis of liquid biopsies emerges as an attractive opportunity for it. Contrary to classical oncology, the therapeutic strategy on precision medicine is based on the distinctive molecular characteristics of patients. Thus, the objective is to tailor patient therapy studying biomarker profiles while reducing, as much as possible, the harmful effects on healthy cells. The latter was already done in BC, where there are a selective treatment depending on tumour subtype. Thus, luminal patients are treated with endocrine therapy while HER2 patients will receive anti-HER2 antibodies [16].
In last years, liquid biopsy is being studied as a tool to comprehend tumour evolution in real-time to guide systemic treatment selection for precision medicine. Moreover, it provides information on the genomic profile of a given cancer and an assessment of tumour burden, without the need for invasive procedures. Although the analytical and clinical validity of the liquid biopsy is increasingly evident, clinical trials that incorporate the analysis of tumour-derived material as ctDNA or CTCs are still necessary for clinic decision-making. Thus, the detection of PIK3CA mutations in plasma ctDNA to guide treatment with the PIK3CA inhibitor alpelisib could provide the first example of a clinically useful ctDNA assay in clinic.
A review of published literature was conducted to assess the use of Liquid Biopsy analysing tumour-derived material (ctDNA, CTC and extracellular vesicles) to identify biomarkers that predict resistance in HR+/HER2-MBC patients treated with CDK 4/6 inhibitors plus endocrine therapy.
Inhibition of CDK4/6 leads to G1 arrest. Pharmaceutical companies envisioned the opportunity and designed treatments targeting CDK4/6. The first generation of CDKi was non-specific and had limited efficacy, affinity and considerable toxicity [6], [17], [21]. Currently, computer-aided drug design is used to develop CDK inhibitors with better potency, selectivity and pharmacological properties studying the spatial structure and inhibition activity of CDKs [22], [23]. Palbociclib, ribociclib and abemaciclib are the last generations of CDK inhibitors that target the ATP binding domain of CDK4/6. The chemical structure determines the specificity against cyclin-dependent kinases, having palbociclib and ribociclib more than 100 fold-higher affinities with CDK4/6 while abemaciclib only ~6 fold higher affinity. A more profound understanding of molecular differences is necessary for the precise use of this drugs in the clinic setting, although it was confirmed the comparable efficacy of these inhibitors by increasing Progression Free Survival (PFS) independently of patients features [24]. MBC patients treated with CDKi who had previously been treated with two or more hormonal line treatments had a higher clinical benefit rate and PFS. It was also observed that the response to therapy was independent of RB1 nuclear expression, the Ki-67 index, p16 loss or the amplification of CCND1 in the tumour tissue. Due to synergetic effect between ET and CDKi, clinical trials focused on combination therapy strategies as the first-line setting for treating MBC patients [25], [26].
Phase I/II study of PALOMA 1 trial assed the safety and tolerability of palbociclib plus letrozole as first-line treatment of postmenopausal HR+/HER2-MBC patients [25], [27]. As in previously preclinical studies, these phase II trial showed patients treated with palbociclib plus letrozole had a higher clinical benefit rate and PFS [17], [25], [28].
Besides, the cohort 2 of the clinical trial was therefore selected based on having CCND1 amplification and/or loss of p16 in the primary tumour to seek a possible biomarker of efficacy, without it could be shown [28] .
Palbociclib plus letrozole was approved as first-line treatment in postmenopausal Luminal MBC patients with no prior treatment for the advanced disease after confirmed palbociclib safety and efficacy in Phase III trial (PALOMA 2) [9], [25]. Later on, PALOMA 3 (Phase III) trial assessed the combination therapy palbociclib plus fulvestrant to treat advanced BC patients that had progressed on previous hormone therapies regardless of menopausal status. HR+/HER2-MBC patients treated with the combinatorial therapy had a higher clinical benefit rate and PFS [12], [17], [29].
Next, the combination of ribociclib plus fulvestrant was assessed in MONALEESA-3 (Phase III) trial. It included postmenopausal patients who progress to (neo) adjunvant ET or presented de novo MBC, who were either treatment naive or received ≤ 1 line of prior ET in the advanced disease setting. The results reported significant PFS improvements irrespective of prior ET [8], [30]. In MONALEESA-7 (Phase III) trial was approved the combination of ribociclib plus goserelin and tamoxifen or aromatase inhibitors (AI) in pre-menopausal MBC patients. It was reported PFS and overall survival (OS) was higher in the combinatorial therapy than in the placebo group [30], [32].
The endpoint of MONARCH 1 (Phase II) trial was determine the activity of abemaciclib as single-agent on continuous schedule in HR+/ HER2-MBC patients who progress on ET and/or chemotherapy [33]. The antitumour activity and manageable toxicities encourage the development of numerous trials to investigate abemaciclib in combination with ET in the first-and second-line settings. In MONARCH 2 (Phase III) trial the combination of abemaciclib plus fulvestrant was approved in MBC patients who progressed during (neo)-adjuvant or first-line ET, regardless of menopausal status [34]. The polytherapy improved objective response rate , OS and PFS, mainly in patients with poor prognosis factors such as visceral metastasis and endocrine resistance [34], [35]. Later, the combination of abemaciclib plus a non-steroidal AI was assessed in MONARCH 3 (Phase III) trial. The polytherapy significantly improved PFS and objective response rate with a tolerable safety profile as initial treatment for HR+/HER2− advanced BC [36].
Preclinical cell line studies have revealed some candidate resistance mechanisms such as upregulation of  [37]. However, it remains to be seen whether these mechanisms identified in vitro are clinically relevant in drug-treated patients.

Deciphering resistance mechanisms through Liquid biopsy analysis
The current molecular cancer studies are based on primary tumour biopsies. Despite its extended use, this technique has multiple downsides: invasiveness, no representation of tumour genetic landscape and its inability to perform serial testing [38]. Tumours are heterogeneous and dynamic units that evolve throughout the disease, sometimes conditioned by the selective pressure exerted by the different treatments received. Therefore, primary biopsy data may not provide real information on current molecular characteristics and biopsies from metastases are iatrogenic and may not represent tumour heterogeneity [39]. Besides, in those patients who suffer from metastasis, sometimes tissue biopsies are not always feasible due to inaccessible tumour sites or the impossibility to sample multiple metastatic sites. In the last decade, it has been technically developed liquid biopsy (LB) as a new diagnostic approach to overcome tissue biopsy limitations. It is based on sampling biological fluids from patients (blood, urine, saliva, etc) to analyse tumour material as circulating tumour cells (CTCs), circulating tumour DNA (ctDNA) and tumour-derived extracellular vesicles (EVs) (Figure 2). Additionally, circulating tumour-derived proteins, circulating tumour RNA and tumour-bearing platelets are other components of LB with diagnostic or prognostic potential [40]. These tumour entities allow assessing the heterogeneity of the tumour, track its genomic evolution during treatment and give more information about the biology behind the metastatic development [41], [42]. Therefore, LB is useful to monitor therapy response and detect resistance which in turn, can help oncologist to predict the progression of the disease, the failure of treatment [43] and the selection of personalized therapies.

Circulating tumour DNA (ctDNA) analysis as a promising tumour biomarker
The ctDNA analysis is an attractive minimally invasive opportunity for genomic study and searches for biomarkers in cancer patients. The amount of ctDNA depends not solely on the amount of death cells but also in the metabolism of the tumour, tumour location, vascularization, rate of proliferation, etc. It is released in the blood stream both passively and actively from primary tumours, CTCs and metastasis [44][45][46]. The ctDNA analysis can be used for diagnosis, to select targeted therapy, for detection of residual tumours and metastases, but also to detect clinical progression identifying resistance mutations [45], [46]. Indeed, in NSCLC (Non-small-cell lung carcinoma) patients, ctDNA has been approved for target therapy selection in advance stages and also in early stages (50%) [47].
Currently, a great variety of studies are assessing the clinical utility of ctDNA in HR+/HER2-MBC treated with CDKi plus ET (Table 1). Despite no association between biomarkers and therapy response was detected in primary tumours [25], [28], [29], some alterations in ctDNA were identified. Subclonal mutations in RB1 at end of treatment were detected in 5 % of patients treated with palbociclib or ribociclib plus ET. The clinical prevalence of RB1 mutations in primary BC tumours is rare, but in patients resistant to CDK4/6 inhibitors with prior endocrine therapy is unknown.
These alterations were the result of selective therapy pressure, since were not detected in the ctDNA obtained before exposure to palbociclib or ribociclib. Also, as these mutations were part of a subclonal population, are difficult to detect in tumour derived material [4], [48]. Furthermore, RB1 mutations were only selected in tumours wild-type for ESR1 mutations, which could suggest that RB1 mutations could be selected when fulvestrant efficacy is not compromised by ESR1 mutation, proposing divergent routes to resistance [48]. in ctDNA abundance was also observed after two weeks of therapy, but did not improve PFS and did not predict sensitivity [52]. Besides, it was observed patients with ESR1 mutations at baseline had a worse PFS than those with wild type mutations, probably due to differentially mutation sensitivity to therapy. Then, it was noticed that ESR1 early clonal dynamics could predict clonal composition at relapse. Loss of ESR1 mutation at end of treatment was more frequent in patients on palbociclib and fulvestrant than those on fulvestrant and placebo. Another study found substantial ESR1 loss and gain through treatment reflects individual tumour subclones as well as the loss of sensible subclones due to therapy pressure [4]. Further analyses are required to comprehend the role of ESR1 mutations in polytherapy resistance. Regarding PI3K alterations, it was described as a gene with a strong pattern of variant acquisition and loss of relatively few clones on treatment [4], [29]. There was no-association between PI3K alterations and PFS, the benefit of the combinatorial therapy or HR status [29]. However, it was observed that a reduction in PIK3CA ctDNA level after two weeks of treatment predicted long term clinical outcome (4 months vs 11 months) [52].  [53]. In other studies, molecular alterations in ctDNA were analysed to identify biomarkers that predict ribociclib plus letrozole therapy response. This polytherapy improved PFS regardless of ctDNA genetic alterations at baseline: PIK3CA, TP53, CDH1, FGFR1, cell cycle-related genes or genes involved in receptor tyrosine kinase signalling [54]. However, Neven et al. found shorter PFS was correlated with altered gene status irrespective of treatment (polytherapy or letrozole alone) [54], [55].
In conclusion, it was observed acquisition of mutations at end of treatment was related with longer PFS in patients who progress later on the polytherapy. It is likely tumours that progress early does not acquire mutations due to lack of treatment pressure. Then, other mechanisms of resistance may dominate in early progression, so it is important to consider intrinsic resistance to select the next line of treatment. In addition, it was described fulvestrant as a major genetic driver of resistance to combinatorial therapy. One possible explanation is tumours are able to adapt to CDK4/6 inhibitors if ESR1 signalling is not correctly suppressed [4].
Currently, an active randomized phase III trial (PADA-1) aims to evaluate the efficacy of switching hormone therapy (from AI to fulvestrant) combined with palbociclib, assessing ESR1 mutations in ctDNA. Likewise, it will determine the safety of hormone therapy and palbociclib. Thus, ESR1 mutations (E380, L536, Y537 and D538 hotspots) will be monitored at baseline and after each cycle of treatment by ddPCR [56]. As preliminary results, ESR1 mutations were uncommon in patients no-treated with AI in the neoadjuvant setting. Besides, one-month treatment with palbociclib and AI decreases ESR1 mutation rate [57], [58].

Circulating Tumour Cells, a possible biomarker to manage HR+/HER2-MBC patients
Cancer heterogeneity results in tumour cells subpopulations that have different rates of proliferation, aggressiveness and drug sensitivity. These cancer tumour cells are released into the blood circulation actively by epithelial-mesenchymal transition or passively detached from primary tumour or metastasis as single cells or clusters, which have a higher metastatic potential. Thus, the presence in blood of ≥ 5 Circulating Tumour cells (CTCs) per 7.5 mL were associated with poor outcome in metastatic breast and prostate cancer, while ≥ 3 CTCs per 7.5 mL in colorectal cancer patients [44]. CellSearch® system (Menarini Silicon Biosystems, Inc) is the only platform validated by FDA for CTC detection and enumeration. It is an immunomagnetic enrichment method that used epithelial antibodies, EpCAM and cytokeratins 8, 18 and/or 19, to positively enrich CTCs. Nevertheless, it ignores CTC subpopulations with mesenchymal or stemness phenotype [44]. Despite the technological advances, the low number of CTCs in the blood system is still a hindrance to their characterization [59]. Thus, sampling higher volumes of blood by leukapheresis is an alternative that is recently being explored. Also, studies at the single-cell level are shedding light on tumourigenesis therapy response or even mechanism of resistance in BC and others tumour types [44], [60].
The single CTC analysis unravels the heterogeneity of the tumour and makes it possible to study resistant clones

Extracellular vesicles and resistance mechanisms to CDKi
Exosomes are extracellular vesicles with a diameter between 50-100 nm that can be released from many cell types.
Among its functions, exosomes have been linked to important roles in cancer biology such as tumorigenesis, angiogenesis, invasion and metastasis. It is also described in the literature that exosomes can transmit drug resistance through functional proteins and microRNAs (miRNA) [62]. . Although all the analysis were performed on primary tumours, this work supports that the study of exosomes could be useful to monitor the response to treatment and the related resistance mechanisms.

Conclusions
The liquid biopsy is being a fundamental tool to study tumour heterogeneity, the main cause of therapeutic failure in cancer patients. Liquid biopsy provides an insight into the dynamic molecular profile of the primary tumour and its metastasis in a non-invasive and real-time approach [54]. A great variety of trials have demonstrated the benefits of combined CDK4/6 inhibitors plus endocrine therapy in HR+/HER2-MBC, such as increasing PFS regardless of menopausal status, prior therapies, endocrine sensitivity and type of metastasis [24], [57]. However, certain limitations were not resolved such as lack of predictive biomarkers to select patients or to detect resistance [25] and is one of the current topics in the context of luminal metastatic breast cancer.
In this regard, studies carried out in ctDNA point to the appearance of subclonal mutations in ESR1 or at the end of polytherapy. Besides, some patients with endocrine resistance are sensitive to CDKi regardless of ESR1 status. This suggests that fulvestrant could be a resistance driver, but expanded clinical studies are needed in this regard. Other proposed biomarkers as FGFR or PI3K lead to contradictory data that do not allow to reach reliable conclusions.
Despite the enormous potential of CTCs, there are no reported works of CTCs gene expression in these patients.
Deciphering changes in expression after combined therapy, especially in resistant CTCs, may be a milestone that allows interpreting the underlying resistance mechanisms. This would be of special interest since those CTCs that bypass therapy, can colonize distal organs and contribute to the progression of the disease. In this sense, we have preliminary data that suggest that the analysis of CTCs can contribute in a relevant way. Concerning exosomes, some publications point to them as very promising biomarkers in the field, but it is necessary to continue with these investigations.
One limitation of studing patients treated with a combined therapy is the lack of knowledge about the contribution of each treatment or if the resistance is due to the action of both drugs. Furthermore, in the clinical context, it remains to be defined whether the mutational state prior therapy determine therapy efficacy . Besides, it should be noted that in tumours that progress early, driver gene mutations evolution is infrequent, probably due to the lack of selective pressure. Therefore, these patients are of special interest due to intrinsic resistance.
Owing to the genetic complexity of cancer and possible mechanisms of acquired resistance, simple models of genetically encoded sensitivity do not reflect the patients' genetic landscape [6]. Hence, is not seeking a standard treatment sequence, but to know the profile of each patient at a certain time to adapt the most beneficial therapy, which means precision medicine.
For all this, the future outlook should be based on molecular characterisation of the primary tumour, metastasis as well as tumour-derived material (ctDNA, CTC or EVs) at different time points in the metastatic clinical setting.
Comprehensive Liquid biopsy analysis of tumour material will change the current clinical paradigm of luminal BC patients completely. Several ongoing clinical trials consider the study of ctDNA but do not include other circulating tumoural entities yet. However, the clinical implementation of liquid biopsy is underway and despite current technological limitations, it is a matter of time before it uses is universal [25], [44], [46].