Effectiveness of Jiedu granule (解毒颗粒) on gut microbiota in patients with advanced hepatocellular carcinoma: a randomized controlled trial

Abstract

OBJECTIVE:

To observe whether Jiedu granule (解毒颗粒, JDG) modulates the composition of the gut microbiota during advanced hepatocellular carcinoma (HCC).

METHODS:

A randomized controlled trial was conducted. Sixty-two advanced HCC participants were randomly allocated to receive JDG or placebo. The median overall survival (OS) times of patients and the variation of relative abundance of bacteria over time were used as main outcome measures.

RESULTS:

Patients who received JDG demonstrated significantly longer median OS times compared with the placebo group. Pyrosequencing of the V3 regions of 16S rRNA genes revealed a dose dependent deviation of gut microbiota in response to JDG treatment. Redundancy analysis identified Clostridium XI and Peptostre-ptococcaceae which related to the onset of liver cancer disappeared after 1-month and 2-month JDG treatment, while in control group, no significant changes of these two bacteria were found. The variation tendency of relative abundance of Bacteroides (essential in immunoblocking therapy of tumor) in JDG group was not obvious while in control group, it was decreased significantly with time. The relative abundance of Roseburia (correlated with the occurrence of liver cancer) was increased in JDG group and was decreased in control group over time.

CONCLUSION:

Changes in the gut microbiota may be associated with the efficiency of JDG on survival period of advanced HCC patients.

Trial registration:

Chinese Clinical TRIAL Registry ChiCTR-OOC-16008002.

School of Traditional Chinese Medicine, Navy Medical University, Shanghai 200433, China

1. INTRODUCTION

In China, liver cancer ranked third in terms of the incidence and mortality estimated in 2015.¹ Hepatocellular carcinoma (HCC) is one of the major histological types of liver cancer, accounting for almost 70%-85%.² Although the advent of targeted drugs such as sorafenib and regorafenib brings hope to advanced HCC patients, the prognosis of advanced HCC remains poor,³ especially for patients with severe cirrhosis. Emerging data draw a microbiome signature that relates to markers of liver disease severity.^4,⇓,⇓-7 More and more studies are now trying to define HCC microbial and metabolite signatures,^8,9 while animal studies have showed the gut microbiota and its metabolites in HCC pathogenesis correlated with peripheral and intrahepatic immune responses.^10,⇓-12 Thereby focusing on gut microbiota alteration might provide a potential therapy for HCC.

For a long time, Traditional Chinese Medicine (TCM) has been an important force for HCC treatment, especially for advanced liver cancer. Years of studies of combination of TCM and Western medicine has provided many Chinese patent drugs, plant extracts, and empirical drugs for clinical use, and their curative effects have been gradually recognized.^{13,⇓,⇓-16} In 2011, Ministry of Health of the People’s Republic of China issued the Standards for Diagnosis and Treatment of Primary Liver Cancer which explicitly added TCM into the systematic treatments of liver cancer, and it pointed out that TCM can improve cancer-related symptoms and quality of life and may prolong overall survival (OS).¹⁷ Jiedu granule (解毒颗粒, JDG), which is a Chinese patent drug, has definite curative effects after more than ten years of clinical validation and is widely used in the prevention and treatment of HCC in our hospital. Previous studies have proved that JDG may prolong the OS and improve the quality of life of HCC patients.^18,19 The mechanism underlying JDG’s impact on HCC has barely been elucidated.

Chinese herbal medicine is usually effective after oral administration and gastrointestinal absorption. In-creasing evidence suggests that gut microbiota plays a crucial role in Chinese herbal medicine therapy by complicated interplay with herbal medicine com-ponents.^{20,⇓,⇓-23} It has been reported that gut microbiota biotransforming herbal medicine chemicals into meta-bolites that harbor different bioavailability and bioa-ctivity/toxicity from their precursors. Herbal medicine chemicals improving the composition of gut microbiota, consequently ameliorating its dysfunction as well as associated pathological conditions.²⁴ Gut microbiota mediating the interactions (synergistic and antagonistic) between the multiple chemicals in herbal medicines. We wonder if JDG works in HCC treatment by gut microbiota.

As a result, we conducted a randomized, placebo-controlled clinical trial to evaluate the efficacy of JDG in the treatment of advanced HCC. Furthermore, we examined the structural alterations of gut microbiota in response to JDG treatment intended to improve OS.

2. MATERIALS AND METHODS

2.1. Study design

A randomized controlled trial (RCT) was conducted in Changhai Hospital, Shanghai, China from March to November, 2017. This trial was registered and was approved by Ethics Committee of Integrative Medicine Institution, Changhai Hospital (number: ChiECRCT-20150073).

2.2. Participants

The study subjects were voluntary inpatients diagnosed as Barcelona clinic liver cancer stage C (BCLC-C) primary liver cancer (PLC) from the Department of Integrative Oncology from Changhai Hospital, Shanghai, China. The diagnostic criteria referred to the Standards for Diagnosis and Treatment of Primary Liver Cancer, 2017.²⁵ PLC patients were included if they met with the following criteria: (a) ≥ 18 years old; (b) clinical stage was BCLC-C; and (c) the expected survival period was more than 12 weeks. While the exclusion criteria were: (a) long-term constipation; (b) having other malignant tumors in the past or present; (c) suffering from cardiovascular diseases; (d) patients with severe infectious diseases; (e) infected with human immunodeficiency virus; (f) gastrointestinal hemorrhage within 30 d; (g) being pregnant; and (h) with language communication barrier.

2.3. Randomization and intervention

All the participants were randomly allocated to two groups according to the random number table method to avoid selection bias.²⁶ Each patient who was enrolled in would receive an opaque envelope containing a random number. Then our oncologist would open the envelope and decided which group would the patient allocated to. To ensure the same sample size of subjects in both groups, remainders would be adopted for further grouping. In the intervention group (JDG group), the patients would be treated with JDG (Tianjiang Pharmaceutical Factory, Jiangsu, China; Production License No. Su ZzY20010266) twice a day for two months while the patients in control group would not take any JDG. The active ingredients of JDG are as follows: root of root of Maorenshen (Radix Actinidiae Macrospermae), root of Shijianchuan (Herba Salviae Chinensis), bulb of Shancigu (Pseudobulbus Cremastrae), and Jineijin (Endothelium Coreneum Gigeriae Galli) (1:1:0.4:0.4).

2.4. Sample collection

Fecal samples should be collected from the patients in both groups for 3 times. For JDG group, the first sample was before the treatment, the second was one month after taking JDG and the third was two months after the treatment. While for the control group, there is a one-month interval between sample collections. All the samples were collected by patients through a Fecalpro Kit (Ruiyi Biotechnology Co., Ltd., Shanghai, China) and within 2 h, the samples would be aliquoted by our oncologists and would be frozen at –80 ℃ immediately.

2.5. DNA extraction and PCR amplification

Microbial deoxyribonucleic acid (DNA) was extracted from fecal samples using the DNA extraction Kit (QIAamp Fast DNA Stool Mini Kit, QIAGEN, Duesseldorf, Germany) according to manufacturer’s protocols. The V3-V4 region of the bacteria 16S ribosomal RNA genes were amplified by polymerase chain reaction (PCR, 95 °C for 3 min, followed by 30 cycles at 98 °C for 20 s, 58 °C for 15 s, and 72 °C for 20 s and a final extension at 72 °C for 5 min) using primers 341F 5′-CCTACGGGRSGCAGCAG)-3′ and 806R 5′-GGACTACVVGGGTATCTAATC-3′. PCR reactions were performed in 30 μL mixture containing 15 μL of 2 × KAPA Library Amplification ReadyMix, 1 μL of each primer (10 μM), 50 ng of template DNA and ddH2O.^27,28

Illumina MiSeq PE250 sequencing

Amplicons were extracted from 2% agarose gels and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) according to the manufacturer’s instructions and quantified using Qubit^®2.0 (Invitrogen Corporation, Carlsbad, CA, USA). After preparation of library, these tags were sequenced on MiSeq platform (Illumina Technology, San Diego, CA, USA) for paired end reads of 250 bp, which were overlapped on their 3 ends for concatenation into original longer tags. DNA extraction, Library construction and sequencing were conducted at Realbio Genomics Institute (Shanghai, China).^27,28

2.6. Process of sequencing data

Tags, trimmed of barcodes and primers, were further checked on their rest lengths and average base quality. 16S tags were restricted between 220 and 500 bp such that the average Phred score of bases was no worse than 20 (Q20) and no more than 3 ambiguous N. The copy number of tags was enumetated and redundancy of repeated tags was removed. Only the tags with frequency more than 1, which tend to be more reliable, were clustered into Operational Taxonomic Units (OTUs), each of which had a representative tag. OTUs were clustered with 97% similarity using UPARSE (http://drive5.com/uparse/) and chimeric sequences were identified and removed using Userach (version 7.0). Each representative tags was assigned to a taxa by RDP Classifer (http://rdp.cme.msu.edu/) against the RDP database (http://rdp.cme.msu.edu/) using confidence threshold of 0.8. OTU profling table and alpha/beta diversity analyses were also achieved by python scripts of Qiime.^27,28

2.7. UHPLC-Q-TOF/MS analysis

Chromatographic grade acetonitrile and methanol were purchased from Merck (Darmstadt, Germany). Chromatographic experiments were carried out on an Agilent 1290 Infinity UHPLC system (Palo Alto, CA, USA), the injection volume was 4 µL. RPLC separation was achieved with Waters XSELECT^TM HSS T3 column (2.5 μm, 100 mm × 2.1 mm, Milford, MA, USA) held at 30 °C. The mobile phase was consisted of solvent A (0.1% formic acid in water, v/v) and solvent B (0.1% formic acid in acetonitrile, v/v). The column was eluted with a gradient of 5% B 0-2 min, 5%-95% B 2-13 min, and 95% B 13-15 min, and the flow rate was 0.4 mL/min. Agilent 6530 QTOF/MS (Palo Alto, CA, USA) was used to perform mass spectrometric detection. Both positive electrospray ionization interface (ESI⁺) and negative electrospray ionization interface (ESI^-) were used for each chromatography separation. Capillary voltage is 3.5 kV (ESI⁺)/3.0 kV (ESI^-), fragmentor voltage is 100.0 V, drying gas flow is 11.0 L/min, and gas temperature is 350 °C. The mass scanning range was acquired from 100 to 1100 m/z over a run time of 15 min.

2.8. Outcome measures

The primary outcome measure of this TRIAL was the variation tendency of relative abundance of bacteria over time between the two groups. For JDG group, the time points for the 3 collection of fecal samples were defined as E0 (before taking JDG), E30 and E60 (one month and two months after treatment); while for the control group, they were defined as F0 (first collection), F30 and F60 (one month and two month later). The secondary outcome measure was the median OS times of patients.

2.9. Follow-up

All the patients were followed by our oncologists by cell phone and would visit again at one-month interval until the patients died.

2.10. Questionnaire

Since diet and drug application would affect the gut microbiota to some extent, to minimize the interference of other factors, the patients would be asked some questions and then our oncologists would fill in a form correlated with their diet and drug application during the past week (Tables S1, S2).

2.11. Diet and medicine

By investigating the diet and drug application of the patients, we found the situation was almost the same between the two groups. More than 80% of patients eat vegetables and eggs as their staple food, and nearly 20% of them ate pork (P > 0.05). Nearly 60% patients took glutathione and compound glycyrrhizin tablets as hepatoprotective drugs and entecavir as antiviral drugs over a long period (P > 0.05). There was no use of antibiotics, gastrointestinal motility drugs, prebiotics, probiotics, targeted drugs and chemotherapeutic drugs (Table S3). All the patients were treated with cinobufagin (intravenously guttae) during their hospitalization. To minimize the influence of environment, each fecal sample was collected on the last day of the patient’s 7-day hospitalization. Meanwhile, all patients were kept the similar diet and environment during hospitalization.

2.12. Statistical analysis

All measurement data were expressed as mean ± standard deviation. All the baseline information of the patients in two groups was processed by student-t test and χ² test (Fisher’s exact method was used if necessary). Since fecal samples should be collected from each patient for 3 consecutive times in a one-month interval, analysis of variance of repeated measurement data was used to test the difference between the two groups. Kruskal-Wallis test was used to analyze the difference of OTU levels in bacteria. A P-value < 0.05 was considered to be statistically different. The statistical analysis mentioned above was processed by SPSS version 21.0 (IBM Corp., Armonk, NY, USA).

3. RESULTS

3.1. UHPLC-Q-TOF/MS of JDG and extract

The representative total ion chromatograms (both ESI⁺ and ESI^- mode) of the granule and extract were shown in Figure S1.

3.2. Study flow and baseline information of the patients

From March to November, 2017, 62 PLC patients (BCLC-C) were enrolled in this trial and among them, 22 were excluded and the rest 40 were randomly allocated to JDG group and control group. Each group had 4 patients lost to follow-up, who had different reasons to quit, such as out of contact or no more willing to go on the trial. Five patients of JDG group and 7 patients of control group were dead within 2 months. Finally, each group had 10 patients finished collection of fecal samples for 3 times (Figure 1).

Figure 1. Flow diagram of this trial.

Homogeneity analysis was performed between two groups from age, tumor size, alpha fetoprotein (AFP) and hepatitis B infection. The mean age of the patients in the two groups were (53 ± 12) and (49 ± 8) years, respectively. No significant differences were found in age (P = 0.43), gender (P > 0.05), tumor size (P = 0.84), AFP (P = 0.98) and hepatitis B infection (P > 0.05) between two groups (Table 1).

Table 1.

Characteristics of HCC patients in JDG group and control group

Characteristic	JDG group (n = 10)	Control group (n = 10)	P value
Age	53±12	49±8	0.43
Sex [n (%)]			> 0.05
Male	10 (100)	9 (90)
Female	0	1 (10)
Tumor size	6±5	7±4	0.84
AFP	581±783	589±760	0.98
Infected with HBV [n (%)]			> 0.05
Yes	9 (90)	9 (90)
No	1 (10)	1 (10)

Notes: data were processed by student-t test and chi-square test. In JDG group, the patients would be treated with JDG twice a day for two months while the patients in control group would not take any JDG. JDG: Jiedu granule; HCC: hepatocellular carcinoma; AFP: alpha fetoprotein; HBV: hepatitis B virus. Difference was considered statistically significant when P < 0.05.

3.3. JDG effectively prolong the survival times of patients with advanced HCC

During a median follow-up period of 10.5 months (range 5-33 months), 8 patients (80%) died and 2 patients survived in JDG group, while in control group, the median follow-up period is 4.5 months (range 4-12 months), and no patients survived. JDG treatment group experienced longer average OS (13.8 months) than control group (5.8 months; P < 0.006) (Figure 2). No drug-related serious adverse events occurred in this study. Finally, JDG can effectively prolong the survival times of patient with advanced HCC.

Figure 2. OS of the patients in JDG and control group.

In JDG group, the patients would be treated with JDG twice a day for two months while the patients in control group would not take any JDG. JDG: Jiedu granule; OS: overall survival.

3.4. Alteration of gut microbiota signature by JDG

Through LEfSe analysis, Clostridium XI and Pepto-streptococcaceae in patients of JDG group showed high abundance before the treatment and low expression after taking JDG for two months (by comparison of E0, E30 and E60, Figure S2). Percentage of relative abundance of Clostridium XI and Peptostreptococcaceae in JDG group, which are correlated with incidence of HCC, showed a significant decreasing tendency over time compared with those in control group (Tables 2, 3, P = 0.034, P = 0.044).

Table 2.

Variation tendency of Clostridium XI

Time	JDG group (n = 10)	Control group (n = 10)	F value				P value
			Time	Time × group		Time		Time × group
Day 0	0.17±0.22	0.12±0.19	3.726	4.232		0.034		0.303
Day 30	0.09±0.16	0.11±0.20
Day 60	0.02±0.04	0.08±0.10

Notes: analysis of variance of repeated measurement data was used to test the difference of variation tendency of Clostridium XI in the two groups. Values are expressed as mean ± standard deviation. In JDG group, the patients would be treated with JDG twice a day for two months while the patients in control group would not take any JDG. JDG: Jiedu granule.

Table 3.

Variation tendency of Peptostreptococcaceae

Time	JDG group (n = 10)	Control group (n = 10)	F value				P value
			Time	Time × group		Time		Time × group
Day 0	0.14±0.18	0.11±0.18	3.423	1.225		0.044		0.306
Day 30	0.06±0.08	0.10±0.19
Day 60	0.02±0.03	0.08±0.10

Notes: analysis of variance of repeated measurement data was used to test the difference of variation tendency of Peptostreptococcaceae in the two groups. Values are expressed as mean ± standard deviation. In JDG group, the patients would be treated with JDG twice a day for two months while the patients in control group would not take any JDG. JDG: Jiedu granule.

So dose the Prevotella, which relative abundance was also decreased in JDG group, and increased in control group over time (Figure 3D). But the difference was not statistically significant (Table 4).

Figure 3. Variation tendency of relative abundance of bacteria on genus level.

A: genus lever barplot; B-D: the variation tendency of Bacteroides, Roseburia and Prevotella over time in the two groups, respectively. JDG: Jiedu granule.

Table 4.

Variation tendency of Provotella

Time	JDG group (n = 10)	Control group (n = 10)	F value				P value
			Time	Time × group		Time		Time × group
Day 0	24±25	4±8	0.014	2.893		0.986		0.068
Day 30	19±28	11±18
Day 60	14±22	14±21

Notes: analysis of variance of repeated measurement data was used to test the difference of variation tendency of Provotella in the two groups. In JDG group, the patients would be treated with JDG twice a day for two months while the patients in control group would not take any JDG. Values are expressed as mean ± standard deviation. JDG: Jiedu granule.

For Bacteroides, which can enhance the efficacy of immune checkpoint inhibitors for HCC therapy, there was no significant trend change in JDG group, while in control group, the relative abundance of Bacteroides decreased significantly with time (Figure 3B). The difference between the two groups at each time point was statistically significant (P < 0.001), and there was an interaction between time and group (P = 0.032, Figure 3B, Table 5).

Table 5.

Variation tendency of Bacteroides

Time	JDG group (n = 10)	Control group (n = 10)	F value				P value
			Time	Time × group		Time		Time × group
Day 0	26±17	42±23	12.397	4.232		< 0.001		0.032
Day 30	28±22	32±19
Day 60	24±21	20±16

Notes: analysis of variance of repeated measurement data was used to test the difference of variation tendency of Bacteroides in the two groups. Values are expressed as mean ± standard deviation. In JDG group, the patients would be treated with JDG twice a day for two months while the patients in control group would not take any JDG. JDG: Jiedu granule. Since P-value of Mauchly’s test of sphericity was < 0.05, data was corrected by Greenhouse-Geisser method.

The relative abundance of Roseburia which is reported to inhibit proliferation of colon cancer cells, was increased in JDG group, while it was decreased in control group over time. There was an interaction between time and group (P = 0.029, Figure 3C, Table 6).

Table 6.

Variation tendency of Roseburia

Time	JDG group (n = 10)	Control group (n = 10)	F value				P value
			Time	Time × group		Time		Time × group
Day 0	1.5±1.7	4.2±5.0	0.326	4.024		0.725		0.029
Day 30	2.8±2.4	4.0±5.4
Day 60	5.8±5.5	1.9±1.5

Notes: analysis of variance of repeated measurement data was used to test the difference of variation tendency of Roseburia in the two groups. Values are expressed as mean ± standard deviation. In JDG group, the patients would be treated with JDG twice a day for two months while the patients in control group would not take any JDG. JDG: Jiedu granule.

In JDG group, the varation of relative abundance of Lachnospiraceae which associated with liver diseases was not obvious over time while in control group, the relative abundance of Lachnospiraceae increased significantly (Figure 4). There was an interaction between time and group (P = 0.044, Table 7).

Figure 4. Variation tendency of relative abundance of bacteria on family level.

A: family lever barplot; B represents the variation tendency of Lachnospiraceae over time in the two groups. JDG: Jiedu granule.

Table 7.

Variation tendency of Lachnospiraceae

Time	JDG group (n = 10)	Control group (n = 10)	F value				P value
			Time	Time × group		Time		Time × group
Day 0	18±18	17±6	0.360	3.508		0.701		0.044
Day 30	14±8	23±12
Day 60	11±9	31±20

Notes: analysis of variance of repeated measurement data was used to test the difference of variation tendency of Lachnospiraceae in the two groups. Values are expressed as mean ± standard deviation. In JDG group, the patients would be treated with JDG twice a day for two months while the patients in control group would not take any JDG. JDG: Jiedu granule.

4. DISCUSSION

JDG displayed a slightly better effect in the survival of advanced HCC patients as no patients survived after a median follow-up of 4.5 months in the control group. Meanwhile, an alteration of microbial signature induced by JDG was found.

Some reports have shown that diet, environment, drugs, prebiotics and probiotics could modulate the gut microbiota for tumor treatment.^29,⇓-31 For example, Newman et al ³² found that mediterranean diet has an anti-breast cancer effect via gut microbiota; Liu et al³³ reported that envrionment factor may be of great significance in recovery from colorectal cancer by increasing brain-gut peptide expression. Chinese herbs had been reported to modulate the gut microbiota structure in type 2 diabetes.²⁴ For the potential effects of Chinese herbs on (a) metabolizing into active metabolites by the action of gut microbiota, (b) regulation of gut microbiota balance, and (c) regulating the fermentation products of the gut microbes, focusing on gut microbiota signature may elucidate their therapeutic mechanisms of cancer therapy.

The strong correlation between gut microbiota and cancer in our study had been documented. It is found that Clostridium XI led to high level of deoxycholic acid (DCA), which activated the MAPK pathway by activating the epidermal growth factor receptor (EGFR) in hepatocytes, or activated the EGFR/Ras/Raf-1/ERK pathway and up-regulate mucin-2 to cause HCC and intestinal cancer.^34,⇓-36 Another research by Ahn et al ³⁷ also indicated that Peptostreptococcaceae is related to the incidence of colorectal cancer. More importantly, Clostridium XI belongs to the family Peptostrepto-coccaceae.³⁸ By LEfSe analysis, we found that the the expression of Clostridium XI and Peptostreptococcaceae significantly decreased after JDG treatment for 60 d. While no significant changes of Clostridium XI and Peptostreptococcaceae were found in the control group after 60 d. Therefore, we considered that JDG might play roles in antitumor by inhibition of Clostridium XI and Peptostreptococcaceae.

Prevotella was a kind of bacterial genus that accounted for a large proportion in the two groups of patients in our study. Studies from Sobhani et al ³⁹ and Akin and Tözün⁴⁰ have shown that the abundance of Prevotella in patients with colorectal cancer is higher than that in healthy people. In our study, Prevotella slightly increased in control group over time but decreased after JDG treatment. However, the difference between the groups was not statistically significant. A larger sample size is needed for future research.

Another bacterial genus accounted for a considerable proportion of the two groups was Bacteroides, which was about 20%-40% (Figure 3B). Nowadays, more and more people begin to pay attention to the role of gut microbiota in the immunomodulation of HCC. Vétizou et al ⁴¹ found that tumors in antibiotic-treated or germ-free mice had no response to lymphocyte-associated antigen 4 (CTLA-4) blockade and their team concluded that gut microbiota is required for the anticancer effects of CTLA-4 blockade. Through further studies, Vétizou et al ⁴¹ discovered that Bacteroides, such as Bacteroides fragilis, can significantly enhance the efficacy of CTLA-4 antibody.³³ Another research by Sivan et al ⁴² found that oral administration of Bifidobacterium alone improved antitumor effect through PD-1/PD-L1 (checkpoint blockade). Back to our study, there was a significant decrease of Bacteroides in the control group after 60 d, while JDG kept the level of Bacteroides unchanged. It was inferred that JDG might benifit the immune checkpoint blocking therapy for HCC by keeping the expression of Bacteroides. However, this hypothesis should be further verified.

Roseburia is a kind of butyric acid-producing bacteria. Butyric is a short chain fatty acid and is a metabolite of undigested and unabsorbed carbohydrates produced by intestinal bacteria.^43,44 It has been already confirmed that butyric can inhibit the proliferation of colon cancer cells and the decrease of Roseburia is correlated with the occurrence of liver cancer.⁴⁵ From the data of our study, we thought that JDG had a positive effect on Roseburia which can improve the antitumor effect of liver cancer.

Lachnospiraceae is thought to be associated with both health and diseases. Mondot et al ⁴⁶ found that the relative abundance of Lachnospiraceae in patients with recurrent Crohn’s disease who had undergone surgery 6 months ago was significantly increased. Moreover, Vacaa et al ⁴⁷ found that patients who enriched with Lachnospiraceae were associated with the occurrence of liver disease such as non-alcoholic fatty liver disease. However, it is also reported that Lachnospiraceae can prevent and treat colon cancer by producing butyric acid.⁴⁸ From the data of our study, we can see in JDG group, the varation of relative abundance of Lachnospiraceae was not obvious over time while in control group, the relative abundance of Lachnospiraceae increased significantly. Therefore, we consider that JDG can inhibit the growth of Lachnospiraceae to some extent.

Our study has some limitations. First, 20 patients withdrew from the trial halfway. Among them, 12 died and 8 gave up. The drop-out rate reached 20%. The reason for high drop-out rate is that HCC patients in BCLC-C stage progress rapidly and complications like jaundice and ascites caused by HCC are difficult to deal with, leading to short survival time of the patients.⁴⁹ During the follow-up process, our oncologists even were told that some patients had lost the ability to get out of bed. Second, cancer patients in China have less active attitude on clinical trials of complementary and alternative medicine compares to patients in Western countries and their attitudes are affected by family members.⁵⁰ More efforts should be made on this in the future.

In summary, we found that JDG might exert antitumor effects and antitumor immunity by alteration of gut microbiota signature. The bacteria such as Bacteroides, Roseburia, Clostridium XI and Peptostreptococcaceae might be involved in its effects.

5. ACKNOWLEDGEMENTS

We would like to thank Dr. Yi Ruan for the additional improvements to the manuscript.