1. Muscle-related miRNAs’ profile in healthy and breast cancer patients
The commercial healthy sample pool and five healthy plasma samples were used to generate the healthy miRNA profile, and the Ct value was shown in supplementary Figure 1 and the relative expression was shown in Table 2. Generally, the expression of muscle-related miRNAs is low in healthy people, especially miR-208 (data not shown), which is undetected in most samples. To explore the impact of breast cancer itself on target miRNA change, we investigated the level of target miRNAs in the plasma of 29 untreated samples. Generally, breast cancer itself did not alter the expression of miR-451, miR-1, and miR-146; compared to healthy counterparts, the expression of miR-21 increased and the expression of miR-133, miR-155, miR-486 and miR-499 decreased in breast cancer patients. However, none of these changes showed statistical significance, as shown in Table 2.
Table 2 Mean expression of target miRNAs in plasma samples from breast cancer vs healthy (miRNA expression Log10)
|
|
|
|
|
Paired samples
|
Non-paired samples
|
Target miRNA
|
Healthy control Plasma sample (Mean±SD, N=6)
|
All untreated plasma sample (Mean±SD, N=29)
|
Mean Fold change in untreated compared to controls
|
P value
|
Untreated samples (Mean±SD, N=14)
|
Treated samples (Mean±SD, N=14)
|
Fold change
|
P value
|
Untreated samples (Mean±SD, N=15)
|
Treated samples (Mean±SD, N=28)
|
Fold change
|
P value
|
miR-451
|
3.05 (0.69)
|
3.25 (1.33)
|
1.07
|
0.718
|
0.24(1.13)
|
1.07(1.32)
|
4.46
|
0.062
|
3.21(1.53)
|
3.66(1.29)
|
1.14
|
0.320
|
miR-1
|
0.37 (0.85)
|
0.35 (0.60)
|
0.95
|
0.927
|
-0.31(0.60)
|
0.12(0.78)
|
-
|
0.125
|
0.46(0.59)
|
0.53(0.68)
|
1.15
|
0.725
|
miR-21
|
0.05 (0.43)
|
0.51 (0.70)
|
10.2
|
0.174
|
0.30(0.70)
|
1.03(0.4)
|
3.43
|
0.002
|
3.20(0.65)
|
3.14(0.58)
|
0.98
|
0.752
|
miR-126
|
2.79 (0.62)
|
3.10 (0.65)
|
1.11
|
0.295
|
0.21(0.66)
|
0.76(0.73)
|
3.62
|
0.038
|
0.59(1.12)
|
1.17(0.80)
|
1.98
|
0.059
|
miR-133
|
1.01 (0.81)
|
0.60 (1.07)
|
0.59
|
0.375
|
-0.40(1.05)
|
0.26(0.94)
|
-
|
0.053
|
2.97(0.90)
|
2.81(1.07)
|
0.95
|
0.541
|
miR-146
|
2.61 (0.68)
|
2.70 (1.18)
|
1.03
|
0.535
|
-0.21(1.41)
|
0.53(1.3)
|
-
|
0.083
|
1.36(0.81)
|
1.27(0.68)
|
0.93
|
0.715
|
miR-155
|
1.41 (0.67)
|
1.14 (0.97)
|
0.81
|
0.881
|
-0.51(1.10)
|
-0.02(0.96)
|
-
|
0.152
|
0.64(0.66)
|
0.71(0.60)
|
1.11
|
0.739
|
miR-486
|
3.25 (0.61)
|
2.80 (1.37)
|
0.86
|
0.782
|
-0.35(1.37)
|
0.61(1.06)
|
-
|
0.002
|
2.70(1.42)
|
3.51(1.09)
|
1.30
|
0.037
|
miR-499
|
-0.66(0.70)
|
-0.19(0.85)
|
3.74
|
0.342
|
0.50(1.12)
|
0.91(0.72)
|
1.82
|
0.239
|
-0.27(0.62)
|
-0.07(0.64)
|
-
|
0.335
|
The expression of target miRNA was normalized by RNU6B, and log10-transformed. Data was presented mean±SD; The comparison of target miRNA was performed by Student’s t test, paired t test, Wilcoxon Rank test or Mann-Whitney test.
2. The impact of cancer subtypes on miRNA change
Considering the distinct biological characteristics and aggressive behaviours among different breast cancer subtypes, it is imperative to investigate the impact of cancer subtype on miRNA expression in breast cancer. Results showed that miR-486 significantly decreased by 32% in Luminal A compared to healthy controls. miR-133 also had a decreased expression in Luminal A despite insignificant difference (p=0.086); Notably, between groups analysis showed that the expression of miR-133 had a significant difference between Luminal A and Luminal B, and it had a higher level in Luminal B than that in Luminal A (p=0.042), as shown in supplementary table 2 and Figure 1. miR-21 has a 19-fold and 14-fold increased expression in Luminal B and Her 2+ breast cancer, respectively, suggesting that miR-21 may be a promising biomarker for inflammation in these two subtypes. Of note, miR-21 has a 4-fold increased expression in Luminal A and 10-fold in basal breast cancer despite an insignificant difference. Inflammation-related miR-126 and miR-146 increased by 30% in Luminal B, and the statistical difference almost reach to margin. These results indicated that Luminal A had the most impact of muscle-related miRNAs, while Luminal B had the most impact on inflammation-related miRNAs. miR-451 has an increased expression in Luminal B breast cancer. Of note, miR-499 had a higher expression level in Luminal B and basal breast cancer compared to luminal A and Her 2+ counterparts, although no statistical significance was observed, as shown in supplementary table 2 and Figure 1. In summary, miRNA may have different expression pattern in different cancer subtypes.
3. The impact of cancer treatment on miRNA change
To explore the impact of breast cancer treatment on target miRNA change, the expression of target miRNA in 42 treated plasma samples (samples belonged to patient who had already received treatment when sample was collected) were investigated. The analysis of non-paired samples (15 untreated samples and 28 treated sample came from different patients) showed that miR-486 significantly increased by 30% in treated samples compared to that in untreated samples; miR-126 in untreated samples was nearly two times of that in treated sample, and the difference almost reach to the significance level (p=0.059). Besides, the analysis of 14 paired samples (untreated sample and treated sample came from the same patient) is consistent with the results of non-paired samples as miR-486 was significantly increased in treated samples compared to its untreated counterparts. Meanwhile, miR-126 and miR-21, instead of showing any significance in non-paired sample analysis, showed a significant change after treatment in paired samples. Of note, the change of miR-133, miR-451, and miR-146 between paired untreated and treated samples almost reach to significance level, as shown in Table 2. These results suggested that miRNA may be promising biomarker for monitoring muscle mass change and inflammation during cancer treatment.
Considering the variety of cancer treatment regimens breast cancer patients received, it is necessary to explore whether the change caused is specific to a certain treatment regimen. miR-21 and miR-486 was significantly increased in post-surgery samples than that in pre-surgery samples, as shown in Figure 2A. Besides, increased expression of miR-21 and miR-486 was significantly in paired pre-chemotherapy samples compared to their post-chemotherapy counterparts, as shown in Figure 2B; such an increased trend was observed in non-paired samples but not statistically significant, as shown in Figure 2C. Notably, except endocrine therapy, miR-21 had an increased trend in other three subtypes, suggesting its stable expression pattern in different cancer subtypes, and may be a promising biomarker for cancer treatment. However, none of these miRNAs was significantly responded to endocrine therapy. The change of miR-486 is similar to miR-21 in different treatment strategies. In addition, miR-486 also significantly responded to combined therapy. Of note, 4 of 5 patients whose plasma samples were in combined therapy group had both surgical resection and chemotherapy, and the response of miR-486 to surgery, chemotherapy, and the combination of both were consistent, indicating that miR-486 may also be a promising biomarker for cancer treatment in breast cancer patients.
In addition, the impact of treatment on miRNA change in different subtypes was also investigated. In paired Luminal A group, miR-21 and miR-486 significantly increased compared to that of untreated counterparts as shown in Figure 3A; while in non-paired Luminal A group, miR-133 and miR-486 significantly increased after treatment, as shown in Figure 3B. Although a slight difference was observed in statistically significant miRNAs, the changing trend of miR-21 and miR-133 is consistent in both paired and non-paired analysis. In Luminal B, miR-146 and miR-155 showed slightly increased compared to healthy controls despite no statistical difference, as shown in supplementary Table 2. Nevertheless, their expression significantly decreased after treatment, indicating that miR-146 and miR-155 may be potential biomarkers for inflammation, which are independent on cancer itself, to cancer treatment in Luminal B, as shown in Figure 3C. In Her 2+, miR-133 and miR-155 did not change in untreated compared to healthy controls, as shown in supplementary Table 2. However, they were significantly increased in treated group compared to their untreated counterparts, indicating that miR-133 and miR-155 may be potential, cancer-independent biomarkers to cancer treatment in Her 2+. In basal breast cancer, none of target miRNA showed any difference. Generally, different cancer subtypes may have different treatment regimen, the expression level of target miRNA between untreated and treatment group in different subtypes suggested that target miRNAs may have different response to cancer treatment regimen.
To explore the time course of selected miRNAs during cancer treatment, plasma samples were divided into different groups according to treatment duration. The detailed information on treatment duration was shown in Table 3. Generally, the change of miRNAs reached to their peak within 91 days, and then gradually return to untreated level, as shown in Figure 4. miR-451, miR133, and miR-486 significantly responded to cancer treatment in 0-91 days group, and none of these miRNAs showed significance afterwards. Of note, 4 of these seven samples in 0-91 days group were from patients who received surgery and the other were from patients received neoadjuvant chemotherapy, and this again underscored the important role of miR-486 to cancer treatment.
]Table 3 The distribution of treatment duration (day)
Group
|
N
|
Median
|
Minimum
|
Maximum
|
Pre
|
29
|
7.00
|
-15
|
50
|
0-91
|
7
|
35.00
|
28
|
91
|
92-182
|
16
|
147.00
|
109
|
182
|
183-365
|
16
|
199.50
|
187
|
343
|
>365
|
3
|
707.00
|
700
|
2112
|
Total
|
71
|
91.00
|
-15
|
2112
|
Note: Treatment duration for each sample was calculated from the day of diagnosis to the day of blood sampling as the exact day of treatment was able to be known from the patients’ data. Minus day in this table means the sampling date was ahead of the diagnosis date.
4. The impact of age on miRNA change
Generally, muscle decayed over natural aging, such a natural process may exert influence on miRNA change, especially muscle-related miRNA change. Therefore, it is imperative to explore the impact of age on miRNA change to cancer treatment. The age distribution of patient cohort was shown in Table 2. Among younger breast cancer (age<=50), age did not have any impact on miRNA response to treatment; while for older breast cancer patients (age>50), miR-133 and miR-486 significantly increased in treated groups compared to its untreated counterparts, as shown in Figure 5.
Table 4 The age distribution in untreated and treated group
|
Age
|
N
|
Median
|
Minimum
|
Maximum
|
Untreated
|
<=50
|
15
|
44.00
|
35
|
50
|
>50
|
14
|
74.00
|
51
|
86
|
Treated
|
<=50
|
12
|
45.50
|
24
|
50
|
>50
|
30
|
74.50
|
51
|
88
|
|
Total
|
71
|
59.00
|
24
|
88
|