Ductal tree ablation by local delivery of ethanol prevents tumor formation in an aggressive mouse model of breast cancer

An average-risk woman has a 12.5% lifelong risk of developing breast cancer [1]. For high-risk women, primary prevention strategies include surgical removal of the breasts and/or ovaries and the use of anti-estrogen medications. Despite the evidence-based effectiveness of these interventions, fewer than 30% of high-risk individuals, such as BRCA mutation carriers, opt for bilateral prophylactic mastectomy and fewer than 15% opt for anti-estrogen therapy as their first choice of preventative treatment [2]. The reasons for this choice are highly personal and vary among individuals and communities, but life-changing consequences and severe side effects are contributing factors [2]. For moderate- and, especially, low-risk women, there are even fewer options available to reduce their risk. Therefore, there is a need to develop new strategies for primary prevention with a focus on high-risk individuals, but strategies that can also benefit moderate- and low-risk individuals.

We seek to develop a minimally invasive procedure as an alternative to prophylactic mastectomy by intraductally delivering a cell-killing solution that locally ablates the mammary epithelial cells before they can become malignant. Our approach is informed by a growing body of literature on the use of intraductal (ID) delivery in the clinic for disease detection, such as ductography, and in preclinical and clinical research studies [3‐17]. ID delivery of cytotoxic compounds, selective estrogen receptor modulators, targeted agents, and/or radioactive particles can prevent tumor formation or provide local disease control in preclinical models [4, 9‐17]. The ID delivery of cytotoxic compounds (e.g., fluorouracil, pegylated liposomal doxorubicin, carboplatin) significantly reduced tumor incidence in an N-methyl-N-nitrosourea–induced rat model [4, 10, 12]. Similarly, ID delivery of pegylated liposomal doxorubicin into the mammary glands of the MMTV-Neu mouse model [10] or cisplatin to the WAP-Cre;Brca1^fl/fl;p53^fl/fl mouse model [9] significantly reduced tumor incidence. Intraductal delivery of lipidoid nanoparticles containing siRNAs against Hox1A into the mammary glands of 12-week-old C3(1)-TAg transgenic mice resulted in a significant decrease in the number of tumors relative to the control group after nine weekly treatments, but tumors eventually developed in most treated glands [16]. Independent clinical studies reported a > 88% success rate in ID administration of pegylated liposomal doxorubicin or carboplatin in up to 8 ductal trees per patient [4, 5]. However, there are limitations with these ID approaches for primary prevention in humans: (i) local cytotoxic therapy (e.g., doxorubicin, fluorouracil, and cisplatin) can induce tumors in treated animals [9, 14, 18], a result that has diminished the enthusiasm for its clinical application, and (ii) local hormonal or siRNA-based therapy would require frequent and repeated intraductal injections [10, 16, 17], which makes it impractical for general clinical application.

Our approach circumvents these limitations by ablating in a single injection per ductal tree all its mammary epithelial cells rather than the targeted killing of existing, highly proliferative, pre-malignant, and/or malignant cells. In this preclinical study, we investigated a chemical ablation approach with ethanol (EtOH) as the cell-killing compound. EtOH is a readily available, stable, inexpensive, and safe compound that has long been used in the clinic. Percutaneous injection of EtOH as an ablative agent is used for the treatment of unresectable liver tumors, renal and adrenal neoplasms, pancreatic cystic tumors, and breast pseudoaneurysms [19‐22] and for celiac plexus neurolysis to reduce pain [23]. Intravascular injection of EtOH as a sclerosing agent is used for the treatment of venous malformations and of spider veins and varicose veins [24‐28]. Here, we demonstrate that the entire ductal tree of a mouse can be filled with a solution containing up to 70% EtOH and that such a filled ductal tree can be imaged in vivo by X-ray microcomputed tomography (microCT). Our results indicate that a 70% EtOH solution is more effective than lower concentrations at locally ablating mammary epithelial cells while causing limited collateral damage to adjacent stroma. Our prevention study using the aggressive and multifocal C3(1)-TAg mouse model of breast cancer shows that ID treatment with EtOH significantly delays tumor formation and significantly decreases tumor incidence. This preclinical study provides support for investigating local ablation as a new strategy for primary prevention of breast cancer.

We investigated and recorded short-term and long-term effects of ID 70% EtOH injections on breast biology and mouse physiology and evaluated whether any of these effects and/or experimental requirements of EtOH-based ablative procedure pose an obvious impediment for human translation. The short-term effects of ID EtOH injection were relatively mild when precautions were taken. The two short-term side effects of EtOH injection were ethylic intoxication and skin laceration. Mice injected with 150 μL of 70% EtOH (3 mammary glands) exhibited signs of alcohol intoxication (about 0.4 g/dL of EtOH content in blood) which was minimized by intraperitoneal injection of 5% sucrose in PBS before and after the ID procedure; animals fully recovered within 4 h after the ID procedure. For the injection of more than 3 glands per mouse, a sequential procedure was performed to allow enough recovery time. The risk of alcohol intoxication in women will be much lower; injection of ductal trees in both breasts, assuming 24 main ducts [41, 42] and 2 mL per duct [3, 8], with 70% EtOH will result in less than 0.1 g/dL of EtOH content in blood, which approximates to drinking three glasses of wine (15 oz), and may cause mild impairment. The total volume of 50 mL of 70% EtOH and total EtOH quantity (27.61 g) is comparable to the up to 50 mL of 95–100% EtOH (up to 39.45 g) reported in percutaneous EtOH injection procedures for the treatment of liver tumors [20] and venous malformations [28]. Skin lacerations were observed in some mice due to small leakage of EtOH from the injected duct or from the needle while exiting the nipple. These were topical lacerations due to direct skin contact and not from internal tissue damage. The risk of these lacerations will be much lower by taping the nipple after ductal injection as it is routine practice for ductography [3, 8]. We did not observe any long-term effects of EtOH injection; specifically, none of the mice had open wounds or infection in injected glands nor did they exhibit any overt signs of pain, distress, or discomfort. Histologically, glands injected with 50 μL of 70% EtOH started to heal by 14 days after the injection. We observed that by 1 month there was limited scarring and subsiding inflammation (Fig. 3). Long-term follow-up of glands showed no signs of scarring or inflammation at 18 months after EtOH injection (Fig. 5). Nonetheless, improvement in the EtOH formulation such as addition of ethyl cellulose as gelling agent to limit the outward diffusion of EtOH should be considered to further minimize the collateral tissue damage. The use of ethyl cellulose for this purpose is routine in some sclerosing protocols for the treatment of venous malformation [26, 27] and has also been reported to improve ablative efficacy in percutaneous EtOH injection in a preclinical model of liver cancer [43].

The safety bar for new preventive agents and approaches is very high. Thus, safety is our primary concern when considering translation of this ID procedure. While the International Agency for Research on Cancer considers EtOH in alcoholic beverages to be carcinogenic to humans (https://​monographs.​iarc.​fr/​wp-content/​uploads/​2018/​06/​mono100E-11.​pdf), this conclusion is based on chronic exposure to EtOH. The exact molecular mechanism(s) of how EtOH increases cancer risk have not been completely established; EtOH metabolization into acetaldehyde, a toxic chemical that can cause DNA damage and DNA-protein crosslinking, is considered a main contributing mechanism for EtOH-induced cancer (https://​monographs.​iarc.​fr/​wp-content/​uploads/​2018/​06/​mono100E-11.​pdf). Nonetheless, we are not aware of any reports of iatrogenic cancer linked to clinical uses of EtOH. Acute exposure to EtOH in mice does not cause significant DNA damage [44]. In addition, we have no evidence from our study that EtOH injection promotes or enhances tumor initiation in cancer-prone C3(1)-TAg animals followed up until they met euthanasia criteria (Fig. 4) nor in non-transgenic animals followed up more than 15 months after ID EtOH injection (Fig. 5). Note that breast tumors were observed in non-transgenic mice within a year of exposure to ID chemotherapy or immunoradiotherapy [9, 13, 14, 18].

We need to mention several limitations of our study. While we attempted to inject as many mammary glands as technically possible, most mice were injected in fewer than 6 glands. Thoracic and abdominal glands were more often injected than cervical or inguinal glands (see Additional file 1: Table S1). Other studies showed the feasibility of injecting all 10 glands [10, 16], but in our hands, cervical glands are not often suitable for injection. There is no need to inject all glands to assess whether EtOH can prevent breast cancer in injected glands with statistical methods that we and other groups used [4, 9, 10]. However, glands that were more accessible for injection may not develop tumors at the same rate as the non-injected glands (see Additional file 1: Figure S1; see also reference [45]). We addressed in part this potential bias with a control group of animals injected with PBS in a similar number of glands and locations (see Additional file 1: Figure S1). In these animals, tumors arose with similar latency and tumor incidence in injected and non-injected glands. Tumor latency in 70% EtOH-injected glands was significantly delayed and tumor incidence significantly reduced compared to PBS-injected glands. Intriguingly, tumor incidence was significantly increased in PBS-injected animals, but not tumor latency, compared to untreated control animals (Table 1). Thus, we used the untreated control group as baseline reference for statistical analysis rather than the PBS-injected group. Typically, the tumor burden in non-injected glands was the reason for euthanasia in EtOH-injected animals, which limited the follow-up time in EtOH-injected glands that might have eventually developed breast tumors. Future prevention studies in other mouse models such as genetically engineered MMTV-Neu [10] and WAP-Cre;Brca1^fl/fl;p53^fl/fl [9] with longer latency and/or lower multiplicity and/or inducible models by ID injection of Cre recombinase [46] would be useful to provide an extended follow-up period and determine the general application of the EtOH ablation procedure to tumor formation driven by different molecular alterations and cell of origin. There are technical challenges in consistently and reproducibly achieving ablation of all mammary epithelial cells of an injected ductal tree. The ductal tree is not completely hollow; it may contain proteinaceous secretions and cellular debris that could affect accessibility and diffusion and/or dilute the EtOH concentration. Architectural changes during the estrous cycle may also affect filling of the ductal tree. These changes are more prominent during alveolar growth in diestrus and subsequent alveolar collapse in proestrus [47]. Thus, injection of the same EtOH volume in ductal tree of different mammary glands of the same mouse or in different mice may have not caused the same rate of ablation providing at least partial explanation as to why tumors still developed in EtOH-injected mammary glands. We observed a varying degree of residual epithelial cell structures in some of the injected glands in non-transgenic animals (Figs. 3 and 5). By design, we injected up to 50 μL of ablative solution, and it is likely that some ductal trees were not fully filled and a larger volume could have effectively ablated more or all of the cells. Due to the size and fragility of the mouse nipple, cannulation to deliver a contrast solution to determine the exact volume requirements before filling with the EtOH solution would be extremely challenging. However, these volume measurements and architectural difference per individual ductal tree could be obtained in women by combining a standard ductography procedure with a preparatory solution for flushing each ductal tree as is often done for ductal lavage collection.

We wish to acknowledge that previous preventive studies directly investigated the idea of epithelial cell ablation as treatment or observed epithelial cell ablation as an indirect consequence of treatment. In a chemically induced rat model, ID injection of a suicidal gene adenoviral vector with the intent of ablating proliferating cells of the TEB to prevent tumor formation paradoxically promoted tumor initiation and increased tumor incidence [11]. Despite the high efficiency transduction, high expression of thymidine kinase, and 50–90% ablation rate upon suicidal gene activation by ganciclovir administration, intraductally treated rats developed tumors at shorter latency than control rats exposed to N-methyl-N-nitrosourea [11]. In genetically engineered mouse models, ID treatment with doxorubicin or cisplatin caused partial destruction of the ductal tree and cell ablation [4, 9], but these treatments were not more effective than those that did not cause cell ablation [4].

Given the existing clinical uses of EtOH and relatively straightforward ID injection procedure, EtOH-based ablation protocols could be readily implemented in clinical trials for primary prevention of breast cancer. We envision that this ablative procedure would most closely approximate the cosmetic treatment of venous malformations. Our preclinical procedure is in line with clinical sclerosing therapy for venous malformation in which patients receive treatment under systemic anesthesia followed by 2 days of anti-inflammatory medications such as NSAIDs that may be extended for a few more days to reduce local inflammation and any possible pain [26]. Typically, the sclerosing therapy reduces chronic pain associated with swollen and deformed vasculature in most patients; this intervention can cause short-term pain in a few patients that is easily managed with medication. In contrast, the physiological response to mastectomy includes local and systemic inflammation, and the scale of these responses and the pain management plan depend on the surgical procedure, including musculoskeletal manipulation and tissue advancement for reconstruction. Typical peri-operative pain management includes a combination of analgesics with regional anesthesia, narcotics, benzodiazepines, and anti-inflammatory medications such as NSAIDs over a course of 2–6 weeks. Regrettably, at least 25% of patients suffer from chronic pain or post-mastectomy chronic pain syndrome [48, 49].

We foresee challenges for clinical implementation that will need to be addressed. (i) The EtOH concentration may need to be adjusted for application in humans due to anatomical differences from rodents, such as the size and complexity of their ductal trees; progressive scalability studies in rats [10, 17] or rabbits [50], and larger animal models, such as pigs, will be required before translation to humans. (ii) If the total amount of EtOH is much larger than 50 mL due individual ducts accommodating more than estimated 2 mL as reported for intraductal delivery of chemotherapeutic agents [4, 5], sequential treatment of each breast in separate visits and/or intravenous administration of thiamine and sugar solution may be required to minimize effects of EtOH intoxication. (iii) The typical human breast is composed of 8–12 ductal trees [41, 42]; successful cannulation and injection in each main duct will be needed to preventively treat the whole breast. (iv) Evidence that the entire tree was filled will be required. EtOH could be injected with a radiocontrast agent to visualize ductal tree filling (as shown in Fig. 1 for mice) using existing ductography methods [3, 8] or could be intrinsically labeled for visualization using other imaging modalities such as EtOH-¹⁷O for magnetic resonance imaging [51]. In some cases, hyperplastic or proliferative regeneration may occlude a duct, preventing filling of the entire ductal tree. It may be possible to dilate or clear passage of such a duct by flushing with a preparatory solution. (v) Pathological evidence that all epithelial cells were ablated may be required. This would be feasible in first-in-human trials in women undergoing elective prophylactic mastectomy, from whom tissue samples of the entire breast will be accessible. We would follow a similar experimental design as described for ID administration of chemotherapeutic agents [4, 5]. Going forward, cell ablation would need to be assessed using in vivo imaging and/or limited tissue sampling by core needle biopsy, since pathological assessment of whole-breast tissue would defeat the purpose of this local, minimally invasive procedure. (vi) The amount of pain, scarring, and other complications associated with this procedure will need to be determined and compared to those of a mastectomy. This procedure will need to have a good safety profile if it were to replace mastectomy in a preventive setting. As with mastectomy, this procedure will eliminate the ability of a woman to lactate or breastfeed.

This preclinical study provides support for further investigating local ablation as a new strategy for the primary prevention of breast cancer in high-risk individuals. This study could also stimulate the evaluation of other chemical and/or thermal ablation strategies. This ID procedure could lead to a breakthrough in breast cancer prevention by providing a universal prophylactic intervention, one not only for high-risk individuals but potentially for those with moderate or low risk; by improving the individual’s quality of cancer-free life; and by decreasing the personal and societal costs of breast cancer.