Topical Application of Houttuynia Cordata Thunb Ethanol Extracts Increases Tumor Inltrating CD8 + /Treg Cells Ratio and Inhibits Cutaneous Squamous Cell Carcinoma In Vivo

Houttuynia cordata Thunb (HCT) is a medicinal and edible herb which has benecial effects on various diseases due to its diuretic, anti-inammatory, anti-oxidative, anti-microbial, anti-viral, anti-cancer and anti-diabetic properties. Most of reports of its anti-cancer activity were conducted in vitro, and its effects on cutaneous squamous cell carcinoma (SCC) has not been investigated yet. Using DMBA/TPA induced SCC mice model, we found that topical treatment by HCT, as well as its bioactive ingredient monomer, eciently inhibited tumor growth. Mechanistically, we found tumor inltrating CD4 + , Foxp3 + T regulatory cells (Tregs) were signicantly reduced and CD8 + /Treg cells ratio was largely increased in tumors after HCT treatment. In addition, several chemokines which recruited immune cells were largely reduced when SCC cancer cells were treated by HCT in vitro. Our results demonstrate the therapeutic effects of HCT on cutaneous SCC and indicate it might inhibit cancer through regulating tumor inltrating lymphocytes and the tumor immune microenvironments.

Skin cancer is the most common cancer worldwide [12,13]. Basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) are the most common forms in non-melanoma skin cancer. In comparison with BCC, SCC is more malignant and has potential to recur and metastasize to other tissues and organs [12,13]. The incidence of SCC has been increased over the past decades [12,14]. Most of SCC can be surgically removed. Radiotherapy and chemotherapy might be used as an adjunct to surgery in advanced SCC patients, if surgery is contraindicated, or if tumor is located in cosmetically sensitive area [15]. There is still a lack of safe and effective medicine for treating malignant SCC.
Exposure to ultra-violet radiation is the most common cause of SCC. Other risk factors include exposure to radiation, carcinogenic chemicals, chronic skin ulceration, and immunosuppressive medication [16].
SCC development is a multistep process which consists of DNA mutation, genome instability, epigenetic changes, in ammation, oxidative stress and tumor microenvironment changes [16][17][18][19]. Mice SCC can be induced by two-stage carcinogenesis protocols which employ mutagenic chemical 7, 12diemthylbenz[a]anthracene (DMBA) as initiators and proin ammatory chemical 12-Otetradecanoylphorbol-13-acetate (TPA) as promoters [20]. This mouse model mimics human SCC exposed to UV or other carcinogens. Large-scale whole exon sequencing revealed that many DNA mutations induced by DMBA/TPA are consistent with mutations in human SCC [21]. In mice SCC model, tumors on back skin can be directly visualized, quantitatively measured and therefore traced individually over time. In addition, drug can be topically applied onto skin, therefore better reach the tumor cells to execute its anti-cancer activity. As we know, drugs are metabolized about 10 times faster in mice than human [22]. Systematic application of drugs often shows no or mild activity in mice, owing that the drugs could not reach the therapeutic concentration in blood. Traditional medicine usually uses the whole plant extracts instead of monomer to treat diseases. They often only show mild to moderate therapeutic effects in human, and these effects might not be observed in mice if they are systematically applied.
Thus, chemical induced mice SCC is an ideal model to evaluate the anti-cancer activity of traditional medicine by topical application.
Cancer cells grow in a special microenvironment where tumor associated cells (immune cells, broblast etc.) and cancer cells interact with each other to reach a balance between suppression and tolerance of cancer cells growth. In ltrating immune cells are major constituents of the tumor microenvironment which play important roles in tumor development, invasion, metastasis, and outcome. Among these in ltrating cells, CD8 + effector T cells are capable of killing cancer cells, while CD4 + , Foxp3 + T regulatory cells (Tregs or T reg cells) secrete a variety of immunosuppressive cytokines, which dampen induction and proliferation of effector T cells [23]. Cancer cells alter the normal homeostatic ratio of effector to regulatory T cells and evade immune surveillance. High in ltrating CD8 + /Treg cells ratios is associated with better prognosis and better response to chemotherapy and immunotherapy in various cancer types [24]. HCT is able to reduce tissue in ltrating in ammatory cells. In rat carrageenan-air pouch model, oral administration of supercritical extracts of HCT suppressed carrageenan-induced exudation as well as in ammatory cell in ltration [25]. In in uenza A virus induced acute lung injury mice model, the lungs treated by avonoid glycosides extracted from HCT presented milder in ammatory in ltration [26]. However, it has not been examined whether HCT treatment can alter the tumor in ltrating lymphocytes and tumor immune microenvironment.
In this study, we used mice SCC model to examine anti-cancer activity of HCT and one of its bioactive ingredients sodium new houttuyfonate (SNH). We used chemical DMBA/TPA to induce SCC and then applied ethanol extracts of HCT or SNH onto back skin. In comparison with control group, the SCC growth was signi cantly reduced without affecting mice body weight, indicating low toxic effects of HCT and SNH. Mechanistically, we found tumor in ltrating CD8 + and CD4 + T cells were both reduced after HCT treatment. More signi cantly, CD8 + /Treg cells ratio was largely increased. Our data demonstrate that HCT is a potential drug for treating cutaneous SCC, and indicate it might inhibit tumor growth through regulating tumor immune microenvironment.

Materials And Methods
Animals ICR mice were obtained from Shanghai SLAC Laboratory Animal Co., Ltd and housed under SPF conditions. All experiments involving animals were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of Shanghai University (the protocol number: 2019033).

Preparation of Houttuynia cordata Thunb ethanol extracts
The dried whole plants were grounded into powders, which were then immersed in 95% ethanol (3 ml of 95% ethanol per 1g of powder) with stirring at room temperature overnight. The mixture was then centrifuged at 2800 x g and ltered through Whatman lter paper. Finally, the ethanol extracts were lyophilized by rotary evaporator. The nal solid extracts were stored at − 20 °C. Before use, solid extracts were re-dissolved in organic solvent DMSO.

Mice skin carcinogenesis model and treatment
Two-stage model of skin tumorigenesis was employed to induce skin cancer. Back skin of 6-Week-old female ICR mice were shaved and then applied topically with DMBA (100 nmol/300 mL of acetone) twice weekly for two weeks, followed by TPA (10 nmol/300 μL of DMSO) twice weekly for six weeks. When tumor appeared, tumor length (L) and width (W) were measured twice a week by vernier caliper. The tumor volume (TV) was calculated according to the formula TV = (L × W 2 )/2. The total tumor volume of each mouse is the sum of all of tumors on back skin. When average total tumor volume of each mouse reached ~100 mm 3 , 14 mice with similar total tumor volume were selected and divided into HCT treatment and control groups. The two groups had similar average total tumor volume and distribution.
Thereafter, the mice were treated with HCT ethanol extracts or the control solvent daily by topical application for another three weeks. TPA would not be further applied during HCT treatment.
In addition, SNH was dissolved in 75 °C ddH 2 O as a 100 mg/kg stock solution and stored at 4 °C. When average total tumor volume of each mouse reached ~100 mm 3 , 27 mice with similar total tumor volume were selected and divided into SNH (20 mg/kg), SNH (100 mg/kg) and control groups. The three groups had similar average total tumor volume and distribution. Thereafter, the mice were treated with SNH or the control water daily by topical application for another ve weeks. TPA would not be further applied during SNH treatment.

TPA-induced skin thickening in mice
Back skin of 6-Week-old female ICR mice (n = 6) were shaved and preteated with topical application of HCT ethanol extracts or the control solvent daily for 10 days, followed by TPA (10 nmol/300 μL of DMSO) topical application daily for 3 days. Skin was then harvested for further analysis.

Western blot analysis of tumor samples
After three weeks of treatment, mice from HCT and control groups were euthanized and the tumors with similar volume were harvested by scissors. Tumor samples were grounded in liquid nitrogen and dissolved in RIPA lysates containing protease inhibitors. Samples were then sonicated and centrifugated.
The supernatant was aspirated and mixed with SDS-loading buffer for further analysis. Protein samples (20-35 μg) were resolved on 12% Tris-glycine gels and transferred onto a nitrocellular membrane. After blocking, the membrane was incubated with the primary antibody and then horseradish peroxidase (HRP)-conjugated secondary antibody. After washing, the HRP substrate was added and the chemiluminescence signal was detected by a CCD camera.

Hematoxylin and eosin (H&E) staining
Brie y, tumor samples were embedded in para n, sectioned and dewaxed. After samples were hydrated, they were stained with hematoxylin and eosin. The samples were dehydrated and then applied with Permount/Toluene solution, covered with coverslip and sealed by nail polish. Images were captured using optical microscope.

Immunostaining
Tumor samples were embedded and frozen on dry ice in OCT compound (Sakura Finetek) and then sectioned at 6-9 μm. Sample sections were then xed in 4% PFA and stained with Alexa Fluor conjugated antibodies. Fluorescence images were visualized and captured by uorescent microscopy.

Statistical analysis
Statistical differences of data were evaluated with two-tailed student's T-test by GraphPad Prism. The difference was considered to be signi cant when P < 0.05 (*), P < 0.01 (**) and P < 0.001 (***).
Quantitative data were presented as mean ± standard deviation.

Topical application of HCT ethanol extracts or its bioactive ingredient inhibits tumor growth
To prepare HCT ethanol extracts, we grounded the dry plant into powders, which were immersed in 95% ethanol to extract the soluble contents. The mixture was then centrifuged, ltered, lyophilized and weighted. The yield of ethanol extraction was 2.5% w/w. The product was then resuspended in DMSO to obtain a 41.67 mg/ml concentration, so that the nal dose of 300 ml of drug applied onto each mouse (~25g) would be 500 mg/kg.
To examine HCT anti-cancer activity, we used two-stage model to induce SCC. Back skin of 6-Week-old female ICR mice were shaved and then topically applied with DMBA twice weekly for two weeks, followed by TPA twice weekly for six weeks, the time point when total tumor volume of most of mice reached ~100 mm 3 . Mice with too small or large total tumor volume were excluded, and the left 14 mice with similar total tumor volume were selected and divided into HCT treatment and control groups. Thereafter, the mice were topically applied with 300 ml of HCT ethanol extracts or solvent DMSO as control daily for another three weeks. TPA application would be stopped during drug treatment.
During treatment, the mice of both groups remained in similar good physical condition. The mice body weight of the two groups occasionally uctuated slightly during treatment, but the overall body weight of the mice remained unchanged (Fig. 1a, b). As revealed by tumor growth curves, HCT treatment signi cantly suppressed tumor growth (Fig. 1a, c). At the endpoint of treatment, the average total tumor volume of each mouse in HCT group reached 637 mm 3 , which was much lower than that in control group with 1041 mm 3 (P < 0.020). In addition, tumor number of HCT group was lower than that of control group, indicating HCT also suppressed emergence of new tumors (Fig. 1d and 1f).
One of main bioactive ingredients of HCT is sodium houttuyfonate (SH). Due to the chemical instability of SH, its adduct analogue, sodium new houttuyfonate (SNH), has been synthesized to improve stability [27]. We then examined whether SNH had similar anti-cancer activity against SCC. As described above, mice with total tumor volume of ~100 mm 3 were divided into three groups and treated by 20 mg/kg SNH, 100mg/kg SNH and control solvent, respectively. As revealed by tumor growth curves, SNH treatment also e ciently suppressed tumor growth as well as number of new emerging tumors in a dose dependent manner ( Fig. 1e and 1f). The overall body weight of the three groups of mice showed no difference (data not shown). All of these data and observation demonstrate that topical application of the HCT extracts or its active ingredient e ciently suppresses SCC growth with low/no toxicity to the animals.
HCT treatment does not change known cancer promoting pathways H&E staining showed similar architecture of tumor tissues from HCT and control groups (Fig. 2a). To investigate the molecular mechanism of how HCT inhibits SCC, we examined whether some known cancer promoting pathways were affected by HCT treatment. The protein b-Catenin is essential for skin cancer stem cells maintenance and SCC growth [28]. Western blot analysis showed that b-Catenin in HCT treatment group was statistically lower than that of control (Fig. 2b, c). However, there was huge variation between tumors even in the same group. Comprehensive genomic analysis of DMBA/TPA induced SCC have revealed that the vast majority of SCC possesses mutations in Hras, Kras or Rras2 [21]. Ras genes are often activated and play important roles during early stages of squamous cell carcinoma development [29,30]. However, we didn't observe change of Ras downstream factor-ERK2 protein (Fig. 2b,  c). The drug pump, ATP-binding cassette sub-family G member 2 (ABCG2), is well known as a speci c marker of the "side population" (SP) of the cancer stem cells and is associated with drug resistance. ABCG2 could be controlled by several pathways, including the PI3K/Akt pathway [31]. ABCG2 protein expression varied between tumors, and there was no statistical difference between HCT group and control group (Fig. 2b, c). Since there was huge heterogeneity of protein expression between tumors, either in different mice or in different tumors from the same mice, it was di cult for us to obtain conclusive results. Therefore, we decided not to further pursue identifying signal pathways altered by HCT treatment.
Topical pretreatment of HCT ethanol extracts suppresses skin thickening induced by TPA It has been reported that TPA alone could induce mouse ear skin oedema and this model has been used as a test for anti-in ammatory activity [32,33]. Similar to ear skin, we found back skin swelling/thickening can be induced by TPA. To evaluate anti-in ammatory activity of HCT, mice back skin was pretreated daily with topical application of HCT for 10 days, followed by TPA (10 nmol/300 μL of DMSO) topical application daily for 3 days. The skin was then xed by 4% PFA, embedded in para n, sectioned and followed by H&E staining. The skin thickness was de ned by the distance from epidermis to the underlying fascia (Fig. 3a, 3b). The average skin thickness in HCT group reached 62.56 μm, which was signi cantly lower than that in control group with 84.68 μm (Fig. 3a, b, c). This result strongly indicates that HCT inhibits SCC through reducing skin in ammation during carcinogenesis.

HCT treatment reduces tumor in ltrating T cells
Since HCT is capable of reducing immune cells in ltrated into in ammatory tissues [25,26], we then investigated whether HCT affects tumor microenvironments and examined tumor in ltrating T cells, which are recognized as the main effectors of antitumor immune responses [34]. Western blot analysis showed that the surface protein markers CD4 and CD8 of T cells were largely diminished in HCT treated tumors (Fig. 4a). Immunostaining analysis showed that there were abundant CD8 + cytotoxic T cells and the CD4 + helper T cells in ltrated into cutaneous SCC tissues, especially in the stroma compartment ( Fig.   4b and 4d). HCT treatment reduced both of these two types of T cells. The CD4 + helper T cells number was reduced much more than that of the CD8 + cytotoxic T cells after HCT treatment (Fig. 4b and 4c).

HCT treatment signi cantly increases the CD8 + /Treg cells ratio CD4 + T cells can be subdivided into regulatory T cells (Tregs) and traditional help T cells. Tregs, formerly known as suppressor T cells, suppress induction and proliferation of cytotoxic T cells, T helper cells and
Antigen-Presenting Cells (APCs), therefore are detrimental to anti-tumor immune responses [35]. Tregs express the biomarker transcription factor Foxp3. Immunostaining results showed that most of CD4 + cells in tumors were Foxp3 + Tregs, and that CD4 + , Foxp3 + Tregs number was signi cantly decreased after HCT treatment (Fig. 4d and 4e). In addition, CD8 + /Treg cells ratio increased about 4 folds after HCT treatment (Fig. 4e).

HCT treatment on SCC cells reduces mRNA expression of in ammatory factors
To investigate the molecular mechanism of why skin in ammation as well as tumor in ltrating T cells were reduced after HCT treatment, we examined mRNA expression level of in ammatory factors in SCC cell line A431 treated by HCT in vitro. The real-time quantitative PCR analysis showed that the mRNA expression of IL-1β, IL-6 and TNF-α was only 11%, 8% and 19% of control group, respectively. These results indicate that HCT treatment might reduce the expression of in ammatory factors in skin cells, therefore reduce the recruitment of in ammatory cells, preventing skin tumorigenesis and suppressing SCC growth.

Discussion
In this work, we evaluated the anti-cancer activity of HCT and its bioactive ingredient in vivo using DMBA/TPA induced cutaneous SCC model. We have demonstrated that topical application of HCT reduces tumor in ltrating T lymphocytes, especially Tregs, increases CD8 + /Treg cells ratio and e ciently suppresses tumor growth without obvious toxicity.
DMBA/TPA induced mice SCC is a unique in vivo cancer model, in which tumors on back skin can be directly visualized, quantitatively measured and traced individually over time. Since SCC has all of hallmarks of cancer development, including DNA mutation, genome instability, epigenetic changes, in ammation, oxidative stress and tumor microenvironment changes [16-19], our study not only demonstrates the anti-cancer activity of HCT on SCC in vivo, but more broadly, indicates HCT might have general anti-cancer activity on other cancer types in vivo.
Chemical induced SCC might develop through activating different cancer promoting pathways which are originated from different DNA mutations [21]. Therefore, it is no surprise that there is huge heterogeneity of protein expression levels between tumors, either from different mice or the same mouse. It is di cult to come to a conclusion of which cancer promoting pathways are altered by HCT treatment, unless huge number of tumors are statistically analyzed.
HCT possesses anti-in ammatory activity [1,2], and it is able to reduce in ammatory cell in ltration in different animal in ammation models [25,26]. HCT treatment reverses oxaliplatin-induced neuropathic pain in rat by regulating Th17/Treg balance [36]. Consistent with previous reports, we found tumor in ltrating lymphocytes especially Tregs largely decreases, and CD8 + /Treg cells ratio increases after HCT treatment. Since Tregs dampers anti-cancer immune response through negatively regulating activation of effector T cells, signi cant increase of CD8 + /Treg cells ratio by HCT treatment can at least partially explain why HCT exhibits anti-cancer activity. In addition, change of in ltrating lymphocytes in tumors by HCT treatment indicates HCT might be able to change in ammatory cell in ltration during early stage of cancer development, therefore affects another process of cancer development. This was supported by the data that HCT pretreatment reduced the skin thickening/in ammation induced by TPA alone.
Since we topically applied the drug onto the tumor, it seems unlikely that HCT modulate the whole immune system and then affect the tumor in ltrating immune cells. Instead, our results support that HCT active components might penetrate into tumor tissue and directly impact on the cancer cells and immune cells around locally. Mechanistically, we have found that direct treatment on SCC cells in vitro by HCT could reduce the mRNA expression of IL-1β, IL-6 and TNF-α from SCC cells. This indicates that HCT might prevent skin tumorigenesis and suppress SCC growth through reducing the expression of in ammatory factors secreted from skin cells and therefore reducing the recruitment of in ammatory cells, including T lymphocytes.
HCT possesses activities against in ammation, oxidative stress and DNA mutations. All of these events are interconnected and play important roles in initiation and progression of cutaneous SCC. Although decrease of tumor in ltrating Tregs and increase of CD8+/Treg cells ratio can explain anti-cancer activity of HCT, we cannot exclude that other activities of HCT might also contribute to this anti-cancer effects.
Nevertheless, our study provides another insight into molecular mechanism of HCT anti-cancer activity: it can re-balance the lymphocytes effector and suppressor in ltrated into tumor, therefore modulate the tumor immune microenvironment to counteract cancer cells.
In conclusion, our ndings enhance understanding of molecular mechanism of HCT against tumorigenesis, and may lead to the development of new topical agents from HCT for the treatment of cutaneous SCC and other cancers. Availability of data and materials The data and materials generated in this study are available from the corresponding author upon reasonable request.

Declarations
Competing interests All authors declare no con ict of interest.
Ethics approval and consent to participate This article does not contain any studies with human participants. All experiments involving animals were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of Shanghai University (the protocol number: 2019033).
Informed consent Not applicable.
Consent for publication All authors are consent to publish this study in the journal of Investigational New Drugs.

Figure 1
Houttuynia cordata Thunb ethanol extracts inhibits cutaneous SCC growth. Cutaneous SCC were induced by DMBA/TPA. When total tumor volume of most of mice reached ~100 mm 3 , HCT or solvent DMSO were topically applied onto skin. Mice tumor volume and body weight were measured twice a week. a Representative images of back skin tumors in control and HCT groups. b The tumor growth curves of the HCT group and the control group. N = 7 mice. P = 0.020. c Tumor number of the HCT group and the control group. d Body weight of HCT group and the control group over time. e The tumor growth curves of the SNH (20 mg/kg) group, SNH (100 mg/kg) group and control group. N = 9 mice. P = 0.209 (20 mg/kg SNH group and control group). P = 0.013 (100 mg/kg SNH group and control group). f Tumor number of the SNH (20 mg/kg) group, SNH (100 mg/kg) group and control group Figure 2 HCT treatment doesn't change known cancer promoting pathways. a H&E staining of tumor samples from HCT treatment or control group. b Proteins b-Catenin, ERK2, ABCG in tumors from HCT treatment or control group were analyzed by western blot. Actin was loaded as control. c Statistic quanti cation of western blot results. For b-Catenin, P = 0.020 Figure 3 Pretreatment of HCT ethanol extracts suppresses skin thickening induced by TPA. Back skin of 6-Weekold female ICR mice were pretreated with topical application of HCT ethanol extracts for 10 days, followed by TPA topical application daily for 3 days. The skin was then xed by 4% PFA, embedded in para n, sectioned and followed by H&E staining. a, b Representative H&E staining of back skin from control group and HCT group. The skin thickness was measured by the distance from epidermis to the underlying fascia (yellow arrows). Scale bar, 100 μm. c Statistic quanti cation of skin thickness. All elds in several sections of back skins from each mouse were counted. N = 6 mice. P = 0.0004 Figure 4 HCT treatment reduces tumor in ltrating T lymphocytes and increases the CD8 + /Treg cells ratio. a HCT treatment reduced the expression of CD4 and CD8 proteins in tumor samples, as analyzed by Western blot. b Representative immuno uorescence images of tumor in ltrating CD4 + and CD8 + cells from HCT treatment or control group. Tumors were frozen in OCT compound, sectioned and immunostained with Alexa Fluor 700 conjugated anti-CD4 antibody or Alexa Fluor 488 conjugated anti-CD8 antibody. c Statistic quanti cation of CD4 + and CD8 + cells number from HCT treatment or control group. All elds of CD4 + or CD8 + cells in several sections from each tumor were counted. N = 4 mice. For CD4 + cells, P = 0.058; For CD8 + cells, P = 0.210. Scale bar, 50 mm. d Representative immuno uorescence images of CD4 + and Foxp3 + cells in the tumors from the control and HCT groups. Tumor samples were frozen in OCT compound, sectioned and immunostained with Alexa Fluor 700 conjugated anti-CD4 antibody and Alexa Fluor 488 conjugated anti-Foxp3 antibodies. Scale bar, 50 mm. e Statistic quanti cation of CD4 + , Foxp3 + Tregs number from HCT treatment or control group. All elds of CD4 + , Foxp3 + Tregs in several sections from each tumor were counted. N = 8 mice. P = 0.0002. f ratio of CD8 + effector T cells number to CD4 + , Foxp3 + Tregs number from HCT treatment and control groups