Adjuvant Photodynamic Therapy, Mediated via Topical Versus Systemic Administration of 5-Aminolevulinic Acid for Control of Murine Mammary Tumor after Surgical Resection

Abstract

Treatment de-escalation is sought in the management of precursor lesions of early stage breast cancer, driving the appeal of adjuvant modalities to lumpectomy that reduce toxicity and minimally detract from patient quality of life. We investigate photodynamic therapy (PDT), with the photosensitizing prodrug, 5-aminolevulinic acid (ALA), as adjuvant therapy to complete resection of murine mammary tumor (propagated from TUBO cells). ALA was delivered either systemically (oral, 250 mg kg⁻¹) at 5 h before 632 nm illumination or topically (20% solution) to the resection site at 10 min before light delivery to 135 J cm⁻². Treatment with either oral-ALA-PDT (oALA-PDT) or topical-ALA-PDT (tALA-PDT) to the mammary fat pad after TUBO complete resection (CR) produced long-term tumor control with 90-day complete response rates of 21% and 32%, respectively, compared to control rates of 0–5% in mice receiving only CR. Thus, CR/tALA-PDT was equipotent to CR/oALA-PDT despite ~10-fold lower levels of ALA-induced protoporphyrin XI as photosensitizer after topical versus oral-ALA administration. CR/oALA-PDT produced more vascular damage, greater proportion of tissue-resident neutrophils and stronger inflammation when compared to CR/tALA-PDT. Collectively, these data provide rationale for ongoing investigation of ALA-PDT as adjuvant therapy after lumpectomy for increased probability of local control in the treatment of breast cancer.

INTRODUCTION

Breast cancer is the leading cancer among women worldwide (1). In the United States, breast cancer is the most common nonskin cancer in women, the second most common cause of cancer death overall in women, and will affect 1 in 8 women in their lifetime (2). In 2021, female breast cancer was predicted to compose 15.3% of all new cancer cases and account for 7% of all cancer related deaths in the United States (3).

Breast cancer screening in the United States has changed in recent years to include options for younger women, beginning at age 40 for those with no family history (4,5). Women with a family history of breast cancer or other risk factors or based on their personal preference may begin screening at even younger ages. These changes in screening are due to rises in breast cancer diagnosis in younger women, as well as declines in breast cancer mortality, due to advancements in treatment and early detection (6). This has led to a marked increase in diagnosis of ductal carcinoma in-situ (DCIS), increasing from 7 cases per 100 000 women prior to mammography screening to 56 cases per 100 000 in the mammogram screening era (7). This marked increase in diagnosis of DCIS has also led to concerns for over-diagnosis and overtreatment of DCIS.

Multiple therapies are used in the treatment of breast cancer including surgery, chemotherapy, radiation, targeted biologics and hormone therapies. Radiation clearly reduces the risk for local recurrence after breast-conserving surgery and is associated with a survival benefit in invasive breast cancer (8), but not in women with DCIS (9). Radiation is associated with a reduced risk for local recurrence and invasive local recurrence after DCIS but not reduced mortality (9). Moreover, radiation, particularly in younger women, is itself a risk for subsequent cancers. One school of thought suggests that radiation is better used for “salvage” of recurrence rather than treatment of in-situ disease. Additionally, radiation has significant side effects and can pro-foundly reduce quality of life when associated with breast pain or decrease in good symmetry and cosmesis of the breasts. These considerations motivate efforts toward treatment de-escalation and a careful risk-benefit assessment of radiation treatment, particularly for early disease or precursor lesions such as ductal carcinoma in-situ (DCIS) (10–12). However, most women with DCIS are not good candidates for omission of radiation because of a much higher risk for local recurrence, and invasive local recurrence, with characteristics of high-grade disease (11). But this has been a source of controversy for decades among those who advocate for reduction in overtreatment by a decreased use of radiation in selected women (13). Thus, there would be clinical utility for alternative adjuvant means to treat DCIS after lumpectomy that have similar risk reduction as radiation for subsequent recurrence and invasive breast cancer, but in ways that lead to fewer side effects. Photodynamic therapy (PDT) may provide an option that meets these needs.

In PDT, a photosensitizing drug is excited by light of a specific wavelength to produce tissue-damaging reactive oxygen species that can be employed to treat malignancy or other pathological conditions. FDA approvals include 5-aminolevulinic acid (ALA) as a photosensitizing prodrug that is a precursor for the endogenous photosensitizer protoporphyrin IX (PpIX) in the heme biosynthesis pathway. ALA generates higher levels of PpIX in tumor, often providing for enhanced selectivity of PDT for tumor tissue (14,15). Furthermore, ALA can be administered either locally or systemically, and, in the case of systemic administration, the rapid clearance of PpIX necessitates only a 24–48 h period of photosensitivity precautions. ALA-PDT has been studied in the treatment of many cancers including those of the skin, brain, and head and neck (16–19) and is FDA-approved for the treatment of actinic keratoses (20). This includes a clinical trial of ALA-PDT in combination with electron radiotherapy for treatment of chest wall metastases of breast cancer (21). Recently, Banerjee et al. (22) made the first clinical report of PDT for treatment of primary breast cancer, in this case with the photosensitizer verteporfin.

Here, we assess the intraoperative addition of ALA-PDT to the complete resection of TUBO mammary tumor in an orthotopic murine model. Others have preclinically studied PDT with ALA or other photosensitizers as a primary treatment for breast tumors (23–27); however, to our knowledge, this is the first description of adjuvant PDT to the resection site of a mammary tumor. These studies are intended to model a potential clinical application of lumpectomy followed by adjuvant PDT to the surgical margins. Our studies incorporate the investigation of PDT with both systemic (oral) and local (topical) delivery of ALA. While high failure rates followed complete resection alone, the addition of PDT, using either mode of delivery, resulted in long-term tumor control. These results are promising for further investigations into the study of PDT as an intraoperative adjuvant to lumpectomy in the treatment of breast cancer. A single treatment for breast cancer is of great interest worldwide with the introduction of intraoperative radiation after lumpectomy and large completed trials for early stage breast cancer (28). PDT could combine benefits of a single postlumpectomy intraoperative treatment and the omission of radiation in these early stage women.

MATERIALS AND METHODS

Tumor models.

Studies were conducted in orthotopic models of TUBO tumors, propagated from cultures of cells (kindly provided by Dr. Brian Czerniecki (29)) and maintained at 37 °C in an atmosphere of 5% CO2 in RPMI media supplemented with 10% FCS and 1% penicillin/streptomycin. Female BALB/cAnNCr mice were obtained from Charles River, NCI Grantee facility at 6–7 weeks of age and allowed to acclimate for one week upon arrival in a University of Pennsylvania conventional rodent facility. At 7–8 weeks old, mice were anesthetized with isoflurane/medical air via nosecone (VetEquip anesthesia machine, Pleasanton, CA), shaved and a depilatory applied for the removal of hair in the lower abdominal area. Mice were then injected with a suspension of 1 × 10⁶ TUBO cells and Matrigel (Discovery Labware, Bedford, MA) in the lower right abdominal quadrant (4th nipple region). Mice were separated into treatment groups 6–7 days postinoculation, once tumor volumes reached 80–100 mm³ as measured by calipers. All animal activities were approved by the University of Pennsylvania’s Institution of Animal Care and Use Committee, housed in an AALAC-accredited facility, which adhered to regulations outlined in the Guide for the Care and Use of Laboratory Animals.

PDT.

PDT was performed in conjunction with either oral or topical administration of photosensitizer. Mice were orally administered 5-aminolevulinic acid (ALA, Sigma-Aldrich, St. Louis, MO) via oral gavage at 250 mg kg⁻¹ (12.5 mg mL⁻¹) 5 h before light delivery for PDT. At the time of resection +/− light delivery, mice were induced and maintained with isoflurane and medical air via nose cone, in dorsal recumbency, on a warm water circulating mat. After anesthetic induction, surgical eye lube was applied, Buprenorphine SR (Zoopharm, LLC, Windsor, CO) was injected at 1 mg kg⁻¹ (0.5 mg mL⁻¹) subcutaneously, and lack of pedal reflex was confirmed. The surgical field was prepped using three alternating rounds of Betadine and ethanol with sterile gauze. The mouse was covered with a lightweight sterile drape and a skin incision was created cranio-caudally adjacent to the right fat pad. The tumor was exposed by making a skin flap along the incision and a macroscopic complete resection was performed using a scalpel to gently work the tumor free from the underlying fat pad. The skin flap was kept secured with hemostats and mammary tissue bed exposed for light delivery. Light was delivered through microlens-tipped fibers to the tumor bed by a 632 nm Ceralas Biolitec diode laser. Light delivery was adjusted to a fluence rate of 75 mW cm⁻² via a Labmaster power meter (Coherent, Auburn, CA) for delivery of 135 J cm⁻² total fluence (30 min) over a spot size of 1.1 cm. At the completion of resection and light delivery, the skin incision was closed using a simple interrupted suture pattern (4-Ethilon), 1 ml 0.9% NaCl was administered subcutaneously to each mouse, and mice were allowed to recover. In mice receiving oral-ALA-PDT without complete reception, a skin incision was created, but tumor was left intact and light was delivered onto the tumor surface.

Separate groups of animals received topical exposure to a 20% solution of ALA. These mice underwent tumor resection, then 0.1 ml of 20% topical-ALA solution (dissolved in 1:1 ethanol: 0.9% NaCl) was applied to the tumor bed using a 1 × 1 cm sterile gauze square. The ALA-soaked gauze remained in contact with tissue for 10 min, then the gauze was removed and light was delivered to the tumor bed. Controls for vehicle alone were exposed only to the 1:1 solution of ethanol: 0.9% NaCl (no ALA) applied topically.

Tumor response studies.

Mice were weighed and assessed for the two subsequent days postsurgery, then observed three times weekly with tumors measured (if applicable) and animals weighed until tumors reached 400 mm³ or the study reached 90 days post-treatment without tumor regrowth. Tumor size was measured by calipers and volume was calculated with the formula V = length × width² × 3.14/6.

Spectrofluorometric assay.

Tumor-bearing animals were exposed to either oral or topical ALA as described above. After tumor resection (before PDT would be delivered), the associated mammary fat pad was collected, protected from light and frozen. Fat pad that was naïve of tumor burden was collected from separate animals. For spectrofluorometric assay, frozen fat pads were thawed, and weights recorded. After mincing, tissue was solubilized using Solvable (PerkinElmer, Waltham, MA) in a hot water bath for 4 h. Following incubation, samples were diluted in water and read on a Fluoromax-4 spectrofluorometer (Horiba, Japan) with a scan from 550 to 750 nm following excitation at 405 nm. After the initial scan was obtained, a known concentration of protoporphyrin IX (PpIX) was added to the sample and another scan was taken. Scans were baselined using FluorEssence v3.5 software and peak heights obtained. Tissue concentrations of PpIX were calculated based on increases in peak height after the addition of a known amount of PpIX.

In-vivo imaging.

Mice were imaged 24 h after complete resection +/− PDT for evaluation of vascular shutdown. An injection of IRDye800 PEG (LI-COR, Lincoln, NE) was administered intravenously 15 min before imaging at 1 uM per mouse. Mice were induced and maintained under isoflurane and medical air anesthesia and images were captured on a LI-COR Impulse instrument using white, 700 nm, and 800 nm fields with mice placed in dorsal recumbency. Using Image Studio 4.0 for analysis, regions of interest (ROI) as boxes were drawn over the right mammary fat pad (tumor-bearing, treated) and left mammary fat pad (tumor-naïve, untreated), aligned as optimally as possible with the surgical incision. Treated and untreated ROI were normalized by division of signal in right mammary fat pad by the left mammary fat pad signal. One week prior to imaging, mice were placed on Teklan Global Rodent Diet (2018Sx) to reduce background image noise.

Flow cytometry.

At one day after treatment, fat pad, along with its associated tumor draining lymph node (TDLN), was removed from mice that were treated as follows: complete resection (CR) only, CR plus vehicle, CR followed by topical-ALA-PDT (CR/tALA-PDT) or CR followed by oral-ALA-PDT (CR/oALA-PDT). For the tumor control condition, the tumor was removed along with the fat pad/TDLN. Fat pads/TDLNs were also removed from naïve (untumored) mice for comparison. All tissues were minced then digested using an enzyme cocktail containing collagenases I, II and IV, DNase I and Elastase (Worthington Biochemical Corp, Lakewood, NJ) as described by Quatromoni et. al (30) mixed with serum-free RPMI. Once a single cell suspension was obtained, the samples were stained for viability (Live/Dead Fixable, Invitrogen, Waltham, MA). The samples were subsequently stained for CD3 (17A2, BioLegend, San Diego, CA), CD4 (GK1.5, BD Pharmingen, San Jose, CA), CD8 (53–6.7, BD Pharmingen, San Jose, CA), Ly6G (1A8, Invitrogen, Waltham, MA), CD11b (M1/70, Invitrogen, Waltham, MA) and CD45 (30-F11, Invitrogen, Waltham, MA). Flow cytometry was performed with a BD FACS Canto flow cytometer (BD Biosciences, San Jose, CA). Data were analyzed in FlowJo v10 software by determining both the percentage of single cells that were gated as CD4+ or CD8+ from the CD3+ population, as well as the percentage gated as both CD11b+ and Ly6G+ from the CD3-population. All values are expressed as a percentage of live cells for comparison across experimental and control groups.

Statistics.

Continuous data were summarized using means and standard errors of the mean (SEM). A time-to-event endpoint was defined as the time for tumor to regrow to a volume of 400 mm³. Kaplan-Meier plots were used to describe the time-to-event endpoints and hypothesis tests were based on log-rank tests. Studies of vascular damage were based on paired samples of treated versus untreated fat pads in the same animal. Within-animal differences on the log-transformed data were assessed using paired T-tests. All other tests were rank-based. For flow cytometric studies, for comparisons to control conditions, a global test was carried out using a Kruskal-Wallis rank-based ANOVA, and in cases with P < 0.05 followed by pairwise rank-based tests. Across studies, experimental conditions of topical versus oral-ALA delivery, un-photosensitized versus ALA-exposed, and tumor-naïve versus tumor-bearing were compared using a Mann-Whitney Rank-Sum test. The Type I error rate was set to 0.05 and all hypothesis tests were two-sided. Two strategies were employed to reduce the chance of false rejections of the null hypothesis. First, we reduced the number of hypothesis tests carried out by prespecifying comparisons, in advance of the analysis, of groups of interest. Second, for analyses using a global Kruskal-Wallis test, groupwise comparisons were only made if the global test was significant, and post hoc analyses were adjusted for multiple comparisons using a Dunn procedure. All analyses used Graphpad Prism (version 8).

RESULTS

Treatment with complete resection followed by oral-ALA-PDT can control murine mammary tumors

PDT efficacy in the treatment of TUBO mammary tumors was studied for orally administered ALA combined with complete resection (CR), mimicking the potential clinical application of PDT for adjuvant therapy after lumpectomy. As shown in Figure 1a, neither CR nor PDT as standalone modalities provided lasting tumor control. For mice that underwent CR, median time-to-400 mm³ was 16 days. This was a modest improvement compared to the 13-day median time-to-400 mm³ for mice that were untreated (P < 0.0001). Treatment of intact (unresected) TUBO tumors with oral-ALA followed by illumination (oALA-PDT (intact)) yielded a median time-to-400 mm³ of 14 days, which was not distinguishable from the untreated tumor. In contrast to these standalone treatments, CR followed by oALA-PDT (CR/oALA-PDT) yielded a median time-to-400 mm³ of 27 days (P < 0.0001 compared to CR). At the conclusion of the 90 day postPDT period, these animals exhibited a 21% complete response (no tumor regrowth) rate.

Figure 1.

Tumor response and photosensitizer accumulation in regimens incorporating oral delivery of ALA. (a) Kaplan-Meier plots of time-to-400 mm³ for TUBO-bearing mice treated as controls (tumored/untreated); exposed to oral-ALA (oALA) followed by light delivery for PDT (oALA-PDT (intact tumor)); treated with complete resection (CR) or exposed to oALA followed by CR and light delivery for PDT (CR/oALA-PDT). Tumors were propagated in the mammary fat pad with 12–14 mice in groups without CR and 21–24 mice in groups with CR. Hypothesis tests used a log-rank test. Light delivery was to 135 J cm⁻² (75 mW cm⁻²) at 5 h after oral delivery of 250 mg kg⁻¹ ALA. (b) Spectrofluorometric assay of PpIX accumulation in fat pad, fat pad resected of TUBO tumor or the tumor itself after oral-ALA delivery. Oral-ALA exposure was a 5 h drug-light interval after 250 mg kg⁻¹ delivered by oral gavage. N = 5–9 per group with plotted values of mean and SEM (error bars). Hypothesis tests based on the Mann-Whitney Rank-Sum test

Although an effective regimen, CR/oALA-PDT of the mammary fat pad produced substantial inflammation over the 24 h following treatment. This manifested physically as swelling in the treated tissue area that in some cases progressed to ascites. To inform factors contributing to this inflammatory response, we evaluated PpIX accumulation in the fat pad after oral administration of ALA. Figure 1b shows that oral delivery of ALA resulted in mean (±SEM) PpIX content of 2.45 ± 0.11 ng mg⁻¹ in normal fat pad compared to 0.09 ± 0.067 ng mg⁻¹ in fat pad not exposed to ALA (P = 0.008). With a value of 3.31 ± 0.81 ng mg⁻¹, PpIX level in the mammary fat bed of resected tumor was statistically no different than that of normal fat pad (2.45 ± 0.11 ng mg⁻¹). Notably, the tumor itself did not accumulate greater amounts of PpIX, with mean levels of 2.93 ± 0.84 ng mg⁻¹. Thus, oral delivery of ALA led to PpIX accumulation in fat pad, both normal and tumor-resected, with the resulting high values potentially contributing to the inflammatory response after treatment with oALA-PDT.

Complete tumor resection followed by topical-ALA-PDT also controls murine mammary tumors

As an alternative to oral-ALA delivery, we additionally investigated topical-ALA-PDT (tALA-PDT) in a treatment protocol that combined CR with tALA-PDT (CR/tALA-PDT). The use of a topical drug (in this case 20% ALA) exposed only the tissue of interest to photosensitizer while also allowing for a drug-light interval of only 10 min, making this a topical drug delivery method that is translatable to an intraoperative application. For this experiment, when mice received CR alone the median time-to-400 mm³ was 21 days (Fig. 2a), similar to the response to CR in mice from Fig. 1a. The addition of a topical application of vehicle alone to complete tumor resection (CR/vehicle) did not alter treatment response, yielding a median time-to-400 mm³ of 24 days. Furthermore, light delivery to mammary tissue beds after complete tumor resection and topical vehicle application (CR/vehicle/light) also failed to alter outcome, with a median time-to-400mm³ of 19 days after light delivery. In contrast, CR/tALA-PDT had a median time-to-400 mm³ of 34 days, 13 days longer than CR (P = 0.0013). Moreover, the complete response rate at 90 days was 32% for CR/tALA-PDT compared to 5% for CR.

Figure 2.

Tumor response and photosensitizer accumulation in regimens incorporating topical delivery of ALA. (a) Kaplan-Meier plots of time-to-400 mm³ for TUBO-bearing mice treated with complete resection (CR); CR followed by exposure to vehicle for topical-ALA (CR/vehicle); CR followed by exposure to vehicle and light (CR/vehicle/light); or CR followed by exposure to topical-ALA (tALA) and light delivery for PDT (CR/tALA-PDT). Tumors were propagated in the mammary fat pad with 18–22 mice per group. Hypothesis tests used a log-rank test. Light delivery was to 135 J cm⁻² (75 mW cm⁻²) at 10 min after topical exposure to 20% ALA. (b) Spectrofluorometric assay of PpIX accumulation after topical-ALA delivery to mammary fat pad or fat pad resected of TUBO tumor. Topical-ALA exposure was for 10 min to 20% ALA. N = 5–9 per group with plotted values of mean and SEM (error bars). Hypothesis tests based on the Mann-Whitney Rank-Sum test

We evaluated levels of PpIX in fat pad after topical-ALA administration (Fig. 2b). A 10 min exposure of mammary fat pad to topical ALA produced mean PpIX levels of 0.20 ± 0.03 ng mg⁻¹, compared to 0.05 ± 0.007 in fat pad without ALA exposure (P = 0.008). After topical-ALA exposure, mean PpIX level in the mammary fat bed of resected tumor (0.32 ± 0.13 ng mg⁻¹) was statistically no different than that of normal fat pad (0.20 ± 0.03 ng mg⁻¹). Thus, measurable levels of PpIX could be detected after just 10 min of exposure to topical ALA, but the levels were about 10-fold lower than the values measured after oral-ALA exposure.

CR/oALA-PDT impairs vascular function

PDT with both orally and topically administered ALA could provide tumor control when delivered in conjunction with complete tumor resection (see Figs. 1a and 2a). In those studies, 48% of mice undergoing CR/oALA-PDT demonstrated vascular effects 24 h postPDT, such as ascites and swelling of the treated mammary tissue area. No notable vascular effects occurred in mice treated with CR/tALA-PDT, which led us to investigate whether there was an association between route of photosensitizer administration and extent of vascular damage. Vascular damage was assessed by optical imaging of tissue-localized IRDye800 PEG following its i.v. injection. Imaging included both the PDT-treated and contralateral untreated mammary fat; fluorescent intensity (FI) data were normalized in each animal to reflect the perfusion of treated relative to untreated fat pad (Fig. 3a). Normalized FI (mean ± SEM) in mice with untreated tumors was 1.19 ± 0.04 (P = 0.013), and it was 1.03 ± 0.06 (P = 0.70) in those that underwent CR. Thus, mice demonstrated slightly greater perfusion in the tumored mammary fat pad versus contralateral normal mammary fat pad. However, after CR, both the tumor-naïve and the tumor-resected fat pad demonstrated similar perfusion. CR/tALA-PDT minimally altered perfusion with a normalized FI value of 1.15 ± 0.16 (P = 0.47). In contrast, significant decreases in vascular perfusion were detected in fat pad treated with CR/oALA-PDT relative to untreated fad pad with a normalized FI value of 0.81 ± 0.06 (P = 0.034). Less perfusion after CR/oALA-PDT can be visualized in Fig. 3b, which shows the fluorescent signal in the resected and PDT-treated region to be dimmer than that in untreated fat pad. For comparison, Fig. 3c depicts a mouse treated with CR/tALA-PDT, in which the fluorescent signal in the treated area is similar to that of the untreated fat pad.

Figure 3.

Vascular damage of the mammary fat pad after complete resection (CR) or CR and PDT of TUBO-bearing fat pad. Fat pads were imaged 24 h post-treatment at 15 min after i.v. injection of LI-COR Dye IRDye800 PEG. (a) Quantification of imaging data expressed as the fluorescent intensity (FI) ratio of treated to untreated mammary fat pad in each mouse. Conditions include tumored (untreated); CR; oral-ALA (oALA) followed by CR and light delivery (CR/oALA-PDT); and CR followed by exposure to topical-ALA (tALA) and light delivery (CR/tALA-PDT). N = 5–7 mice per group. Values shown are means and SEM for the paired values (ratios). Hypothesis tests of whether the ratio is equal to 1.0 based on paired t-tests with *P = 0.013 and #P = 0.034. (b) Representative image of treated and untreated mammary fat pad from a mouse that received CR/oALA-PDT; weaker FI is apparent on the treated side. (c) Representative image of treated and untreated mammary fat pad from a mouse that received CR/tALA-PDT; FI is similar on the treated and untreated sides. Dark vertical lines in the images are the stitches from CR

CR/oALA-PDT and CR/tALA-PDT differentially alter mammary fat pad immunological milieu

Flow cytometry was performed to understand the inflammatory and immune cell types associated with oral versus topical-ALA-PDT after a CR (Fig. 4a). First, we determined how the presence of tumor changed the immune cell milieu of the mammary fat pad. As expected, due to the contribution of CD45− tumor cells, tumor-bearing fat pad demonstrated a higher percentage of CD45− cells than tumor-naïve fat pad, with this cell population increasing (mean +/−SEM) from 0.75 ± 0. 12% of total live cells in the naïve fat pad to 12.56 ± 3.55% in tumor-bearing fat pad (P < 0.0001) (Fig. 4b). Moreover, the presence of tumor also altered the relative populations of neutrophils (CD11b+, Ly6G+), CD4+ and CD8+ cells. The percentage of CD11b+, Ly6G+ cells increased in tumor-bearing compared to naïve fat pads (Fig. 4c). Tumor-naïve fat pad contained 0.09 ± 0.03% neutrophils and this value increased to 0.54 ± 0.074% in tumor-bearing fat pad (P < 0.0001), albeit we note that on an absolute basis the Ly6G+ composition of the tumor-bearing fat pad remained low. Conversely, percentages of CD4+ T cells were higher in the tumor-naïve compared to tumor-bearing fat pad with values of 57.15 ± 0.87% compared to 33.00 ± 2.30% (P < 0.0001), respectively (Fig. 4d). Similar trends were observed for CD8+ T cells, with its relative population decreasing from 20.95 ± 1.12% in tumor-naïve fat pad to 12.12 ± 0.79% (P < 0.0001) in the presence of tumor (Fig. 4e).

Figure 4.

Flow cytometry of immune cell milieu in naïve versus tumored mammary fat pad. (a) Schematic of the gating strategy for flow cytometry. Populations of (b) CD45-, (c) CD11b+ Ly6G+, (d) CD4+, and (e) CD8+ cells in either tumor-naïve or TUBO-bearing mammary fat pads, expressed as a percentage of the population of live cells. N = 12–15 samples per group. Values shown are means and SEM. Hypothesis testing based on Mann-Whitney Rank-Sum tests

Treatment of the tumor-bearing mammary fat pad led to further changes in the immune environment. We evaluated the effects of complete resection (CR), performed independently or together with PDT as CR/oALA-PDT or CR/tALA-PDT. Not surprisingly, when compared to tumor-bearing mammary fat pad, the percentage of CD45− cells was lower in all conditions that included CR to remove tumor burden; values ranged from 1.17 ± 0.5% to 5.98 ± 2.31% among treatments that included CR in contrast to 12.56 ± 3.55% in the tumor-bearing fat pad (Fig. 5a). These values reached significance for CR/vehicle and CR/tALA-PDT, showing fewer CD45− cells to be associated with topical treatments. Moreover, across most conditions that utilized complete resection there were significant increases in the number of neutrophils in the fat pad (Fig. 5b). After treatment that incorporated CR, mean percentage of neutrophils increased from 0.53 ± 0.07% in tumor-bearing fat pad to 23.65 ± 12.02%, 20.15 ± 8.48%, and 20.41 ± 6.41% after CR, CR/vehicle, and CR/oALA-PDT, respectively. Importantly, the percentage of neutrophils was insignificantly elevated after CR/tALA-PDT, increasing only to 9.49 ± 7.66%. CD4+ and CD8+ T cell populations did not change among the different conditions that utilized CR (Fig. 5c,d). The CD4+ percentage means ranged from 23.97 ± 6.52% to 37.43 ± 4.67%, while means for CD8+ cell populations ranged from 9.72 ± 2.83% to 13.24 ± 2.89%.

Figure 5.

Flow cytometry of immune cell milieu in TUBO of mammary fat pad treated with complete resection (CR) or CR and PDT. (a) As the percentage of live cells, populations of (a) CD45−, (b) CD11b+ Ly6G+, (c) CD4+, and (d) CD8+ cells in mice that were tumored (untreated); received CR; received CR and the vehicle for topical ALA (CR/vehicle); exposed to oral-ALA (oALA) followed by CR and light delivery for PDT (CR/tALA-PDT); or received CR followed by exposure to topical-ALA (tALA) and light delivery for PDT (CR/tALA-PDT). Mean and SEM (error bars) for n = 6–12 samples per group. A Chi-square-test based on a Kruskal-Wallis ANOVA was used as a global test. If P < 0.05 for the global F-test, pairwise comparisons between the control and each of the active treatments were made using the Dunn test to adjust for multiple comparisons. Conditions of CR/oALA-PDT and CR/tALA-PDT were compared by Mann-Whitney tests

Flow cytometry further revealed specific immunological differences between CR/oALA-PDT and CR/tALA-PDT. The percentage of CD45− cells was significantly lower after CR/tALA-PDT compared to CR/oALA-PDT, suggesting that topical ALA-PDT led to greater tumor cell clearance. Also, topical treatment with ALA-PDT better controlled the neutrophil population than did oral-ALA-PDT after CR, that is, a significantly (P = 0.0176) lower percentage of neutrophils was associated with CR/tALA-PDT to the mammary fat pad (9.49 ± 7.66%) compared to that found after CR/oALA-PDT (20.41 ± 6.41%).

DISCUSSION

The primary finding of this report is that ALA-PDT can provide control of TUBO mammary tumor when delivered to the resection bed after complete removal of tumor from the mammary fat pad. Long-term tumor control was achieved in 21 −32% of animals in which intraoperative PDT was given after a complete resection of all visible disease. Importantly, control was achievable whether ALA was administered systemically (oral) or locally (topical), and in conjunction with an initial set of treatment conditions (drug dose, drug-light intervals and light dose) not yet optimized to maximize treatment response. Thus, these data provide rationale for ongoing investigation of ALA-PDT as adjuvant therapy to lumpectomy for increased probability of local control in the treatment of breast cancer.

Although both topical and oral ALA could introduce tumor control after a complete resection, the characteristics of the PDT response were different as a function of the mode of ALA delivery. Topical administration of ALA may have allowed for more direct exposure of ALA and production of PpIX in the tumor cells that remained after tumor resection, potentially providing for more direct cytotoxicity to these cells than found after systemic ALA delivery. CR/tALA-PDT was also associated with some (mild) neutrophil influx at the 24 h time point, thus a role for productive, controlled inflammatory and innate immune responses cannot be ruled out after CR/tALA-PDT as contributing to its efficacy. However, this response was much more controlled than the visible inflammation that could result from CR/oALA-PDT. In mice treated with a complete resection followed by oral-ALA-PDT, vascular damage to the tumor bed in the mammary fat pad was apparent within 24 h after PDT. In-vivo imaging of PEGylated dye was used to measure PDT-created vascular shutdown (31) showing CR/oALA-PDT to significantly decrease the perfusion of treated mammary fat pad. This vascular damage aligns with expectations because inflammation was visibly observed and elevated neutrophil content was detected in the CR/oALA-PDT treated mammary fat pad. Overall, the extent of vascular shutdown after CR/oALA-PDT was mild (~20%) compared to the more extensive vascular damage that can be associated with PDT, but we emphasize that the tissue under study is the mammary fat pad at the site of tumor resection, not the tumor itself. Since the treatment site consists of microscopic disease in a bed of otherwise normal tissue, it is undesirable to produce a complete shutdown of tissue vasculature as might be sought when PDT is applied to a solid tumor mass. Thus, treatment effect on normal vasculature is contributing to the efficacy of CR/oALA-PDT, but needs to be controlled to avoid excess toxicity.

Physical examination also yielded evidence of a vascular effect after CR/oALA-PDT that can be contrasted to a milder reaction when ALA was delivered topically. Namely, we found the mice treated with CR/oALA-PDT, but not those treated with CR/tALA-PDT, to exhibit ascites and local swelling at 24 h postlight delivery. Interestingly, in work that spans from preclinical to clinical studies, reports on vascular disruption with topical-ALA-PDT vary from evidence of significant blood vessel damage to mild or no evidence of a vascular effect (32–35). In an explanation of these findings, it has been noted that the vascular effects of topical-ALA-PDT associate with PpIX production in the vessel wall, thereby making the extent of vessel damage highly dependent on aspects of the treatment protocol that alter PpIX concentration (36). When a systemic photosensitizer (Photofrin) was employed, treatment of subcutaneous breast cancer tumors resulted in effects similar to the gross observations that we made in our model. Ahn, et al. (37) observed bruising after PDT of EMT-6 breast cancer tumors, with a darkening of the treatment area initiating at 48 h after light was delivered. This was not unlike an edematous discoloration we sometimes observed at 48 h after CR/oALA-PDT and is consistent with the development of vascular damage for these treatment conditions. Of note, any decreases in tissue perfusion during the course of light delivery could reduce the extent of PpIX photobleaching during light delivery, which importantly acts to protect tissue from damage (38).

Flow cytometry was further used to characterize the inflammatory environment of the treatment site, studying the immune cell composition of the mammary fat pad after its exposure to CR/oALA-PDT versus CR/tALA-PDT. Resulting data from 24 h after PDT showed there to be a larger proportion of CD11b+, Ly6G+ cells in the mammary fat pad of CR/oALA-PDT-treated mice than in CR/tALA-PDT-treated mice. These data indicate a higher neutrophil count and suggest a larger inflammatory component to treatment in mice that received oral vs. topical-ALA-PDT. A neutrophil response to ALA-PDT has also been noted in numerous studies by other investigators, including in clinical applications of ALA-PDT to skin. In studies of topical-ALA-PDT of healthy buttock skin, Evangelou et al. (39) reported neutrophil infiltrate to peak within 4 h after PDT, and begin to decrease by 24 h postPDT. After ALA-PDT treatment of actinic keratoses, Nakaseko et al. (40) reported that infiltrates of neutrophils were observed over time points ranging from hours up to 1 week following PDT. In particular, this group observed a pronounced influx of neutrophils in the hours after PDT that persisted through the one day observation time point. However, after 3 days and up to 7 days after PDT, lymphocytes dominated the layers of PDT-treated skin. The observed timing of these immune cell influxes may inform the results of our studies, suggesting that neutrophil levels in CR/tALA-PDT-treated fat pad may have peaked earlier and then normalized by the 24 h time point of our assay. Inversely, 24 h may have been early to find substantial changes in lymphocyte numbers, thereby explaining the minimal PDT effect that we detected on these populations. Ultimately, our ability to detect neutrophil increases at 24 h in fat pad treated with CR/oALA-PDT could be a result of a stronger and longer-lasting inflammatory response to CR/oALA-PDT versus CR/tALA-PDT, a finding that would be consistent with physical observations after each of these treatments.

There also exists controversy over the contribution of neutrophils to the efficacy of ALA-PDT. In preclinical studies of systemically delivered (i.v.) ALA, de Bruijn et al (41) describe a greater influx and larger accumulation of neutrophils in PDT-treated rhabdomyosarcoma that associated with more effective PDT conditions. According to this work, the increase in neutrophils in tumors was maximal at 24 h after treatment. However, those neutrophils were not necessary for an effective immune response because rats given antigranulocyte serum to prevent neutrophil influx did not show a significant decrease in the effectiveness of ALA-PDT as assessed by tumor regrowth. In contrast, in studies of Photofrin-PDT, deVree et al. (42) found the number of neutrophils in circulation to be a determinant of PDT efficacy. In their study, neutrophil-depleted tumors experienced a large influx of neutrophils and subsequent regression in tumor growth after antigranulocyte serum treatment was stopped. Similarly, Kousis et al (43) found neutrophils to be essential in tumor response to PDT mediated by 2-[1-hexyloxyethyl]-2-devinyl pyropheophorbide-a (HPPH). In our studies, it would be valid to question the contribution of neutrophils to outcome (as found by de Bruijn et al (41)) based on findings that tumor response was similar to CR/oALA-PDT and CR/tALA-PDT irrespective of differences in the extent of neutrophil influx that was induced by these conditions (at 24 h postPDT).

Perhaps the most notable difference between CR/oALA-PDT and CR/tALA-PDT in this study was the amount of PpIX accumulation at the resection site as a function of the mode of ALA delivery. As to be expected, compared to topical exposure, systemically administered ALA resulted in higher tissue levels of PpIX (14). These higher PpIX levels are likely a key factor that contributes to inflammation and vascular damage after delivery of oALA-PDT to the resection site. In contrast, PpIX accumulation in the mammary fat pad was much lower after topical exposure to ALA. We found that just a 10 min topical exposure to ALA was sufficient to increase PpIX levels to more than that detected in unphotosensitized fat pad, albeit these levels were ~10 times lower than those produced by oral-ALA delivery. Most impressively, despite the modest levels of photosensitizer relative to oral delivery, CR/t-ALA-PDT achieved tumor response equipotent to CR/oALA-PDT.

This success with short exposure times to topical ALA is not unprecedented. In studies of a 20% solution of topical ALA for the treatment of actinic keratoses, Kaw, et al. (44) tested the effects of blue light illumination immediately after ALA application. Even for the shortest illumination time of 30 min, beginning light delivery immediately after exposure to ALA was equally as efficacious as the longer topical-ALA incubations that were tested. Moreover, shorter ALA incubation resulted in minimal treatment-associated pain. In other studies, Zhang et al. (45) concluded that a 20% topical-ALA solution is as therapeutically successful as systemically administered ALA, if exposure times are short (30 min), and if the photosensitizer was sprayed on the treatment area twice with an interval of 30 min between. Revealingly, this protocol, which involved PDT of the rabbit vocal folds, also resulted in less severe inflammation than that created by PDT after systemic ALA delivery. In additional studies of ALA delivery, this group noted that only 5 min of exposure to topical ALA allowed the drug to penetrate deeper than 1 mm in the vocal folds, while just a 15 min incubation provided for ALA levels in the superficial 1 mm of tissue that were equivalent to those after a 2 h incubation.

Future extensions of this work could further research the development of ALA-PDT as an adjuvant to lumpectomy in treatment of early stage or precursor lesions of breast cancer. Of note, in another approach that combined surgery with PDT, Gao et al (46) reported that surgery followed by topical-ALA-PDT for patients with Extramammary Paget’s Disease resulted in no recurrence for 6 months post-treatment and a 29.03% recurrence rate at 1-year post-treatment for 31 patients. In the treatment of breast cancer, it is important to consider that disease can be diffuse, potentially leaving behind microscopic disease despite the presence of clean margins (47). Recurrences, including as invasive lesions, can develop secondary to undetected invasion at initial diagnosis (48). Thus, ongoing work could consider approaches to ensure optimal and more complete (homogenous) production of PpIX, such as through the introduction of preconditioning regimens as have been well described for breast cancer and other types of premalignant dysplasia and malignant carcinoma (49,24,23).