*2.2. Method Validation*

The method validation was conducted in strict accordance with US Food and Drug Administration (FDA) guidelines [23], the content of which consists of selectivity and specificity, linearity, limit of quantification (LOQ), limit of detection (LOD), accuracy, precision, recovery, matrix effect, and stability.

In the process of method development, a selectivity and specificity test was used to verify that the measured substance is the intended analyte to minimize or avoid interference. The selectivity and specificity test of this experiment was demonstrated by the analysis of blank plasma from six individual rats, which was examined by comparing the retention times of DOX, HER, and IS in blank plasma, the addition of DOX and HER at LOQ and IS to blank plasma, and a plasma sample 2 h after an intravenous administration of the mixture of DOX and HER. The blank plasma should be free of interference at the retention times of the analytes and the IS, and which in spiked samples and actual samples should be consistent, respectively. A typical MRM chromatogram of mixed blank plasma in rats, spiked plasma samples with DOX and HER at LOQ and the IS, and plasma samples of rats after an intravenous injection of a mixture of DOX (5 mg/kg) and HER (5 mg/kg) for 2 h is shown in Figure 2. The results showed that there was no significant endogenous interference in the retention time of the analyte under the established chromatographic conditions.

Calibration curves were established by plotting the peak-area ratio (y) between analytes (DOX or HER) and IS against the nominal concentrations. Linearity was evaluated by weighted (1/*x*2) least squares linear regression analysis. The correlation coefficient (*r*) should be greater than 0.99, indicating a good linearity. The limit of detection (LOD) is defined as the lowest detectable concentration, judged by the signal-to-noise ratio (SNR) >10. The limit of quantification (LOQ) was defined as the lowest concentration on the calibration curve, which represents the sensitivity of the method and should be lower than the minimum concentration in all the samples. The linear calibration curve was obtained by plotting the peak area ratio (analytes/IS) versus DOX and HER concentration. A weighted (1/*x*) quadratic least-square regression analysis gave typical regression curves. The calibration curves, correlation coefficients, detection ranges, and LOQ of DOX and HER in plasma and myocardial tissues are shown in Table 1. The calibration curves had good linearity with the corresponding range of DOX and HER (r > 0.99). Under the optimized conditions, DOX LOQ was <4.0 ng/mL, and HER LOQ was <2.0 ng/mL in rat plasma, judging from the signal-to-noise ratios (SNR) of >10.


**Table 1.** Calibration curves of doxorubicin and hernandezine in plasma and myocardial tissue homogenate of rats.

Accuracy and precision tests are critical in determining whether the method is ready for validation, and involve analyzing replicate quality controls (QCs) at different concentrations throughout the assay range. Specifically, the intraday and interday precisions and accuracies were obtained by analyzing five replicates of QC samples at three levels for three consecutive days. Precision, defined as the relative standard deviation (RSD), should be within 15% at each QC level. Accuracy expressed as relative error (RE) must be within ± 15%. Except for the LOQ level, the RSD value of precision should be within 20%, and the RE value of accuracy should be within ± 20%. The intraday and interday precision of the QC samples of DOX and HER were lower than 9.3% and 5.6%, respectively. The accuracy of DOX was −14.0% to 5.5%, and the accuracy of HER was −9.0% to −0.8% (see Table 2). All the assay values were within the range of acceptable variables, indicating that the established method was precise and accurate.


**Table 2.** Precision and accuracy of doxorubicin and hernandezine in plasma of rats (n = 5). RSD: relative standard deviation.

Recovery of the analytes should be optimized to ensure that the extraction is efficient and reproducible. Recoveries of the analytes at three QC levels (*n* = 5) were determined by comparing the peak area ratios of the analytes to IS from QC samples with those of analyte solutions spiked

with post-extracted matrix at equivalent concentrations. The matrix effect was examined to assess the possibility of ion suppression or enhancement. The matrix effect was measured by comparing the peak area ratios of the analytes to IS in solutions spiked with the blank processed matrix with the solutions at three QC levels. In common, it was considered that the matrix effect was obvious if the ratio was less than 85% or more than 115%. The recovery and matrix effect data of DOX and HER in rat plasma were shown in Table 3. The matrix effect range of all analytes was 92.9 ± 4.3% to 112.8 ± 1.8%, and the RSD value was lower than 11.6%. The average recovery of DOX and HER at three QC levels was 88.7 ± 6.2% to 108.4 ± 4.9%, and the RSD value was lower than 7.0%. The results showed that this method had no matrix effect, and could be used for biological analysis.


**Table 3.** Matrix effect and recovery of doxorubicin and hernandezine in plasma of rats (n = 5).

Stability was conducted by analyzing three replicates of the samples at three QC levels under the following conditions, including bench top stability after 4 h of exposure at room temperature, auto-sampler stability after 24 h of storage in the auto-sampler at 4 ◦C, freeze/thaw stability evaluated for three freeze–thaw cycles after freezing at −80 ◦C and thawing at room temperature, and long-term stability storage at −80 ◦C for 30 days. The samples were considered stable if the average percentage concentration deviation (expressed as RSD) was within 15% of the actual value. The stability results are shown in Table 4. The variation of all the stability studies was less than 15.0%, which met the standard of stability measurement. Therefore, this method could be used for routine analysis.

**Table 4.** Stability results for doxorubicin and hernandezine in plasma of rats under different storage conditions (n = 3).


#### *2.3. Pharmacokinetics*

This validated method has been successfully applied to the determination of plasma concentration of DOX and HER in rats. In this study, we compared the pharmacokinetic parameters of DOX in the combined treatment group with those in the single treatment group. The pharmacokinetic profiles of HER were also compared in the same way. The mean plasma concentration–time profiles of DOX and HER for the three groups were shown in Figure 3. The pharmacokinetic parameters of DOX and HER in rats following the intravenous administration of single DOX (5 mg/kg), single HER (5 mg/kg), and a combination of DOX and HER (5 mg/kg, respectively) were shown in Table 5.

*Molecules* **2019**, *24*, 3622 of DOX in the combined treatment group with those in the single treatment group. The pharmacokinetic profiles of HER were also compared in the same way. The mean plasma concentration–time profiles of DOX and HER for the three groups were shown in Figure 3. The pharmacokinetic parameters of DOX and HER in rats following the intravenous administration of single DOX (5 mg/kg), single HER (5 mg/kg), and a combination of DOX and HER (5 mg/kg, respectively) were shown in Table 5. 

**Figure 3.** (**A**) Mean plasma concentration–time curves of doxorubicin in a single doxorubicin group and combination group; (**B**) Mean plasma concentration–time curves of hernandezine in a single hernandezine group and combination group. **Figure 3.** (**A**) Mean plasma concentration–time curves of doxorubicin in a single doxorubicin group and combination group; (**B**) Mean plasma concentration–time curves of hernandezine in a single hernandezine group and combination group. **Table 5.** Non-compartmental pharmacokinetic parameters of hernandezine and doxorubicin in a single doxorubicin group, single hernandezine group, and combination group (n = 6).


The Cmax of DOX in the single group and combined group was 2647 ± 650 ng/mL and 5703 ± 2980 ng/mL, respectively. Meanwhile, the AUC0–∞ was 1412 ± 114 ng/mL and 2453 ± 218 ng/mL, respectively. Meanwhile, the t1/2 was 4.6 ± 0.8 h and 4.2 ± 0.6 h, and the MRT0–∞ was 4.9 ± 0.9 h and 4.5 ± 0.5 h, respectively. Significant differences of Cmax and AUC0–∞ of DOX were observed between the single and combined groups with equivalent doses of DOX administration, which indicated that HER could increase the absorption of DOX. However, there was no significant difference between the t1/2 and MRT0–∞ of DOX, which indicated that HER had no effect on DOX's elimination and excretion. In turn, we could see from the plasma concentration–time curves of HER in two treatment groups in Figure 3 that the combination use of DOX made the pharmacokinetic behavior of HER no longer fitted to a non-compartmental model that was used to calculate the pharmacokinetic characteristics in this study. However, we were still able to reach a conclusion from the plasma concentration–time curve and the pharmacokinetic characteristic of HER that the free drug concentration of HER was reduced by the combination use of DOX. The possible reason might be the enhancement of DOX on the drug–protein binding of HER.

The comparison of the accumulated concentrations of DOX in myocardial tissue 8 h after intravenous administration of single DOX and combination of DOX and HER was investigated as shown in Figure 4. A significant difference between the two groups could be observed (*p* < 0.05), indicating that HER was able to reduce the accumulation of DOX in myocardial tissue. Meanwhile, recent studies demonstrated that doxorubicinol (DOX-ol), a secondary alcohol metabolite of DOX [24,25], which may have caused cardiac toxicity by being poorly cleared from the heart and accumulating there to form a long-lived toxicant to heart [26], was to blame. Therefore, the next step is to study whether HER could inhibit the conversion of DOX into DOX-ol, which might be considered as a therapeutic target for DOX-induced cardiac toxicity. *Molecules* **2019**, *24*, x 9 of 13 

**Figure 4.** The comparison of the accumulated concentrations of doxorubicin in myocardial tissues 8 h after the intravenous administration of doxorubicin and doxorubicin + hernandezine (mean ± SD, *<sup>n</sup>* <sup>=</sup> 6, *p* < 0.05). **Figure 4.** The comparison of the accumulated concentrations of doxorubicin in myocardial tissues 8 h after the intravenous administration of doxorubicin and doxorubicin + hernandezine (mean ± SD, *n* = 6, *p* < 0.05).
