Photodynamic therapy reduces metastasis of breast cancer by minimizing circulating tumor cells

: Cancer metastasis after traditional surgery introduces a high barrier to therapy efficacy. Photodynamic therapy (PDT) for cancer is based on a photochemical process of photosensitizers that concentrate in tumors and release oxidant species under light excitation to destroy cells. Compared with traditional surgery, PDT provides minimal invasion and targeted therapy. In this in vivo study, we monitor the real-time and long-term dynamics of circulating tumor cells (CTCs) after a single round of PDT and after surgical resection in a breast cancer animal model. The CTC level is low after PDT treatment, and the recurrence of the primary tumor is postponed in the PDT group compared with the resection group. We find that metastasis is correlated with the CTC level, and the PDT-treated mice show no metastasis in the lung or liver. Our results suggest PDT can effectively reduce metastasis by minimizing CTCs after treatment and is a great technology for breast cancer therapy.


Introduction
Photodynamic therapy (PDT) has been widely used in clinical cancer therapy [1][2][3]. The PDT strategy takes advantage of light and chemical photosensitizer molecules that localize and concentrate in tumors and release high-concentration oxidant species (free radicals, singlet oxygen, triplet species) under light excitation to kill cancer cells for targeted cancer therapy. The clinically approved photosensitizers for PDT, including aminolevulinic acid (ALA) and porphyrin, can accumulate in tumors due to the abnormally high metabolism of cancer cells, although the specificity might be limited [3][4][5][6]. The photosensitizers are then excited by specific light illumination, which is focused on the tumor to minimize side effects and off-target damage. The light is usually delivered by direct free-space illumination or endoscopic systems to excite the accumulated photosensitizers (PSs) in the tumor [7][8][9]. The tumor cells undergo apoptosis or necrosis after PDT, while the normal cells nearby suffer little damage. In this way, the PDT achieves a minimally invasive anti-tumor effect. Compared with traditional surgery, PDT may effectively protect patients from invasive physical harm to the body.
Tumor metastasis has been recognized as a key factor inducing the high risk of death during cancer development and after therapy [10]. Metastasis is a complex process that mainly consists of the following events: (A) some tumor cells in the primary tumor acquire an aggressive phenotype; (B) those tumor cells invade the intercellular matrix and move toward blood vessels, usually accompanied by epithelial/mesenchymal transitions; (C) some tumor cells migrate into blood vessels and survive in the circulating systems, at which point they are called circulating tumor cells (CTCs); (D) some CTCs occasionally exit the blood vessels and enter distant organs/tissues; (E) those CTCs can only survive if they evade the innate immune response; and (F) the surviving CTCs have a chance to develop metastatic deposits if they proliferate actively under the microenvironment [10][11][12]. In general, CTCs are believed to have an essential role in tumor metastasis.
Traditional surgery imposes the risk of residual tumor cells entering the circulatory system through the incisions made in blood vessels or other physiological structural changes and thus becoming CTCs [13][14][15]. In addition, the great damage to the body by surgical operation and other traditional treatments, including radiotherapy and chemotherapy, may activate the development of cancer stem cells [16][17][18]. Therefore, therapeutic regimens to treat the primary tumor may initiate metastasis. In contrast, PDT brings little physical invasion or off-target damage effects. In this study, we report that the PDT strategy can effectively reduce metastases in the breast cancer mouse model by minimizing the CTCs, encouraging the clinical implementation of PDT as an effective strategy against metastasis.

Animal model
MDA-MB-231 cells (human triple-negative breast cancer epithelial cell line) were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured in Dulbecco's modification of Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 10,000 U/mL penicillin, and 10,000 µg/mL streptomycin. MDA-MB-231 cells transfected with stable expression of mCherry were cultured in DMEM with 10% fetal bovine serum (FBS), 10,000 U/mL penicillin, and 10,000 µg/mL streptomycin in Petri dishes. Using conventional in vitro flow cytometry (FACS Aria II, Becton, Dickinson and Company, Franklin Lakes, NJ, USA), the MDA-MB-231-mCherry cells with higher fluorescence intensity were selected to establish tumor mouse models. The tumor volume was evaluated by the following formula: long diameter × short diameter 2 )/2.
Female nude mice aged 5-6 weeks and weighing 20 ± 2 g were purchased from Shanghai Lingchang Experimental Animal Co., LTD (Shanghai, China) and maintained under standard housing conditions in the Department of Experimental Animals, School of Biomedical Engineering, Shanghai Jiao Tong University. Models of transplantable orthotopic tumors were established in the nude mice by injecting MDA-MB-231-mCherry cells into the abdominal fat pad of the nude mice. The concentration of MDA-MB-231-mCherry cells was set at 2 × 10 7 /mL in 50 µL PBS mixed with 50 µL Matrigel (#354262; Corning, NY, USA) for orthotopic tumor implantation. Therefore, in this study, all CTCs emitted red fluorescence of mCherry under excitation. All efforts were made to reduce animal suffering. All animal experiments were performed in accordance with guidelines evaluated and approved by the Ethical Committee of Animal Experiments, School of Biomedical Engineering, Shanghai Jiao Tong University, China.

PDT protocol
The ALA (Sigma, # A3785) was dissolved in phosphate buffered saline (PBS) with the concentration of 25 mg/mL ALA was injected intraperitoneally into tumor-bearing mice at a concentration of 250 mg/kg four hours before PDT. The mice were anesthetized during PDT.
The light source was a customized LED lamp with tunable power density via two convex lenses (GCL-010204, DaHeng OTF). The optical power was measured using a power meter (Thorlabs, PM100D) installed with the sensor head (Thorlabs, S121C, measurement range: 400-1100 nm). The power density was calculated by the detected optical power divided by the effective detection area of the sensor head (70.85 mm 2 ) since the light spot was large enough to cover the whole sensor head. The tumor was irradiated by white light (400-700 nm, 100 mW/cm 2 ) for 20 minutes for a single time. An opaque cloth was used to cover the rest of each mouse's skin to protect it from light.

CTC measurement
We used a previously reported in vivo flow cytometry (IVFC) system to monitor the CTCs in the animal model in vivo [19]. The IVFC system applied a 535 ± 15 nm light-emitting diode (LED) to visualize the major vessels of the ear microcirculation by bright-field microscopy. An artery with a diameter of 50 to 70 µm was selected for detection. A 561 nm laser was collimated, re-shaped, and combined to focus on the sample by an objective (Olympus, Lucplfln 40 ×, N.A. = 0.6, dry) to excite the fluorescence of flowing CTCs genetically labeled with mCherry in the selected vessel. The laser beam was shaped by a cylindrical lens to a slit located transversely on the blood vessel of the mouse. The size of the laser slit at the focal plane was approximately 5 × 72 µm. The depth of focus (i.e., the full width at half maximum of the optical slit on the sample in the axial direction) was about 50 µm, which was selected to match the blood vessel of interest. The fluorescence at 610-670 nm from mCherry of the passing CTCs was collected and coupled into a photomultiplier tube (PMT). In this way, fluorescently labeled CTCs that traveled through this slit could be accurately recorded one by one.
For CTC detection, mice were anesthetized and then stably fixed on the sample stage of the IVFC system. The arteries and veins in the ear of the mice were imaged by bright-field imaging. The slit-shaped laser focal spot was moved to the blood vessel and located transversely to excite mCherry fluorescence of CTCs across the section. The mice were monitored for different durations after being anesthetized. As a control, we checked every mouse for an hour to see if there was any original endogenous red fluorescence in the blood vessels without tumor implantation. The mice were observed weekly for the first three weeks after PDT treatment and every two weeks thereafter.

Statistics
The metastases were quantified by selecting the metastases in the microscopy images of pathological sections from the lung and liver of the mice. Statistical analysis was performed in GraphPad Prism 8, and all error bars displayed on graphs represent the mean ± standard error of mean (SEM) by the one-tailed paired Student's t-test unless stated otherwise. The baselines in the plots of fluorescent signal versus time were normalized to eliminate the background vibrations due to breathing and the heart beating. The fluorescence intensity was acquired as the peak amplitude of the signals.

PDT minimized CTCs in the animal model of breast cancer
To monitor CTC dynamics in real time, we constructed an animal breast cancer model with MDA-MB-231-mCherry cells as described in Methods on Day 0. On Day 15, the mice were treated with PDT (for 20 minutes with ALA) or surgery to eliminate the primary tumor, as shown in Fig. 1(A). To evaluate the tumor recurrence after PDT, the tumor was treated with a single round of PDT. The CTCs were monitored by in vivo flow cytometry (IVFC) on Day 14 (before treatment) and every week from Day 17 (after treatment) to Day 31, and then once every two weeks after that (only the last time measurement was specifically performed on Day 75 (not 73) just before Day 76 to sacrifice the mice). For the control, a group without ALA that received only light illumination and a group without any treatment were observed simultaneously. On Day 76, all mice were sacrificed for anatomy analysis. The metastases were analyzed by using pathological sections from the lung and liver with hematoxylin and eosin (H&E) staining. The tumor volume in each group was recorded, and the results were shown in Fig. 1(B). In general, the tumor volumes in the PDT and resection groups were significantly smaller than those in the light and control groups. Recurrent tumors after resection and single-round PDT were discovered around Day 24 and Day 32, respectively ( Fig. 1(B)). The weight of the mice in each group was shown in Fig. 1(C). Each mouse was detected by the IVFC system for 1 hour in vivo to acquire the CTC dynamics. The typical CTC fluorescence signals of mCherry were shown in Fig. 2(A). The CTC counts could be determined according to the fluorescence peaks (inset in Fig. 2(A)). It could be seen from Fig. 2(B) that the count of CTCs after PDT maintained a significantly lower level compared with the other three groups. Specifically, the count of CTCs after PDT was significantly lower than that after resection, although the tumor volumes of both groups were very close.

PDT reduced the risk of tumor metastasis in lung and liver in the breast tumor model
Since we performed only a single round of PDT on Day 15, which was usually not enough to eliminate all local tumor cells, a recurrent tumor was found on Day 32 (Fig. 3 A and B). In comparison, the recurrence time after resection was much earlier (on Day 24, Fig. 3(B)). Notably, the development of the recurrent tumor in the PDT group was significantly slower than that of the resection group ( Fig. 3(C)), implying a better tumor cell elimination effect by PDT than by resection. The final relative recurrent tumor volume of the PDT group was also smaller. It is important to note that the relative CTC levels on Day 17 (two days after the treatment) and Day 75 in the PDT group, normalized by the CTC level on Day 14 (one day before the treatment), were significantly lower than those in the other three groups. Therefore, we evaluated the metastases in those groups to investigate the correlation between CTCs and metastases. The lung and liver of the mice were dissected for pathological sections with H&E staining to visualize the metastases on Day 76 (Fig. 4(A)). A group of normal mice without tumor implantation was also presented as a negative control. The typical secondary tumor in the lung and liver were exhibited. No metastases were found in the normal or PDT groups. The control group without any treatment exhibited the most metastases. We quantified the metastases by calculating the areas of the secondary tumors (Fig. 4(B)). The light-treated and control groups had the largest areas of metastases in the lung. We then analyzed the correlation between the CTC counts on Day 75 and the metastases in those groups. As shown in Fig. 4(C), the metastasis area was highly correlated with the CTC count. The light-treated and control groups with high CTC counts presented high-level metastasis, whereas the PDT and resection groups showed low CTC counts and metastasis. Interesting, the R 2 coefficient of linear fitting between CTC count and metastasis area was 0.99. The same quantification method was used to analyze the metastases in the liver (Fig. 4(D)). The R 2 coefficient of linear fitting was 0.75 ( Fig. 4(E)). This result suggested that CTCs mediated metastasis during cancer development and PDT reduced metastases in lung and liver of orthotropic breast tumor by minimizing CTCs, even though primary tumor recurrence emerged two weeks after PDT.

Discussion
Tumor metastasis has always been the main reason for the low efficiency of treatment and an important factor in high mortality. CTCs play a key role in metastatic dissemination. A number of studies have reported that CTC dynamics are of great significance in cancer metastasis [12,[20][21][22]. The relationship between the prognosis of breast cancer patients and CTCs has been reported [23,24]. Braun et al. proposed the CTC count in patients with breast cancer correlated with increased metastatic burden, aggressive disease, and a decreased time to relapse [23]. In this study, we present the real-time CTC dynamics after PDT treatment of the primary tumor in the mouse model of orthotropic breast cancer. As shown in Fig. 2 and 3, PDT can reduce the CTC count and shorten the recurrence time of tumor. The reason might be due to the therapeutic effect of PDT, which is a combination of PDT-mediated cytotoxicity and antivascular effects [1,[25][26][27][28] that prevent tumor cells from migrating to the blood. Previous studies have reported some tumor vascular responses to the blanching or vasoconstriction by PDT, including platelet aggregation and tumor angiogenesis [29]. It should be noted that the light irradiation and generation of reactive oxygen species during PDT can block blood vessels by exerting oxidative stress on the blood [30][31][32]. Compared with traditional resection of the tumor, PDT prevents the chance of residual tumor cells leaking into the blood flow through physical incisions, and it introduces little direct damage to the whole body. Although the PDT may lack specificity to tumor cells and cause some side effects, including off-target photodamage to skin or the tissue surface, the minimal invasion can still protect the body [33]. In general, tumor metastasis is highly correlated with CTC count, which can potentially be increased by therapeutic strategies that require incisions into tissue.
It should be noted that in this study, to examine the relationship between CTCs and metastasis, only a single round of PDT with a moderate irradiation dosage was performed so that CTCs could still be observed after PDT, and as expected, tumor recurrence emerged two weeks after this single round PDT. In the clinic, several rounds of PDT are usually performed for a thorough treatment of the tumor. There is still a significant limitation of PDT due to the penetration limit of light in tissue [34]. Usually, the PDT can only work for skin cancer, oral cancer, esophagogastric cancer, and other tumor types with cancer cells mainly in the tissue surface. Some PDT schemes for large-volume tumors have been proposed by using optical fibers that need to be inserted into the tumor tissue. We expect that our study may provide a research basis and shed some light on PDT among a variety of tumor types in the future. Data availability. Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.