Timofey N. KomarovCenter of Pharmaceutical Analytics LLC

The column is crucial

Shim-Pack column enhancement to separate challenging substances

Timofey N. Komarov, Igor E. Shohin, Margarita A. Tokareva, Olga A. Archakova, Dana S. Bogdanova, Alexandra A. Aleshina, Natalia S. Bagaeva, Veronika V. DavydanovaCenter of Pharmaceutical Analytics LLC, 117246, Russia, Moscow

Development of new drugs also requires finding new analytical methods to monitor therapies as well as to control the way a drug enters the body, is absorbed and digested, known as pharmacokinetics. The methods must be labor- and cost-efficient. Interaction of the single components of an analytical system such as the chromatograph and the optimal column, is also important.

HPLC with various detection methods is actively used to determine the content of drugs in biological fluids. One of the most difficult practical tasks is the chromatographic separation of poorly retained compounds – drug substances poorly retained on the chromatographic column. Valganciclovir and Ganciclovir are examples of such substances; both are used against herpes viruses – ganciclovir as an infusion and valganciclovir in various dosage forms.

The aim of this study was to develop a method for determination of valganciclovir and ganciclovir in human plasma by LC-MS/MS for pharmacokinetic and TDM studies. This method was developed and validated. Linearity in plasma sample was achieved in the concentration range of 5-1,000 ng/mL for valganciclovir and 50-10,000 ng/mL for ganciclovir.

HPLC as one of the modern analytical techniques is now widely used to determine the content of medicines in biological liquids. Depending on the chromatographic conditions and detector used, it is possible to separate and identify substances with different properties and, especially, the substances having similar structure.

Challenge: separation of medicines

One of the most challenging practical tasks is the separation of the medicines which are weakly retained on a chromatographic column. The various chromatographic methods could be used for that purpose: HILIC (Hydrophilic interaction liquid chromatography) and ion-exchange columns, ion-pair reagents as eluents etc. However, these approaches may not be effective where a mass spectrometer is used as detector, or when the medicines are measured in complex biological matrices. The case of valganciclovir (VAL, figure 1) and ganciclovir (GAN, figure 2) illustrates this situation.

Figure 1: Chemical structure of valganciclovir
Figure 2: Chemical structure of ganciclovir

Ganciclovir was developed initially as a drug for the treatment of cytomegalovirus infection, after which it was discovered that GAN could inhibit in vitro other herpes viruses (human herpes viruses 1 and 2 types, Epstein-Barr viruses, chickenpox virus, human herpes virus type 6 etc). [1] Typically, ganciclovir was used intravenously only during a therapy, and a new prodrug named valganciclovir was developed to increase the bioavailability. Valganciclovir metabolizes easily into ganciclovir which in turn provides a therapeutic effect. [2, 3]

Like many other antiviral drugs, valganciclovir and ganciclovir are both rather hydrophilic substances, as clearly shown by the values of their octanol/water partition coefficients (log P, table 1). These properties must be considered in the development of analysis methods including sample preparation and separation techniques.

Valganciclovir Ganciclovir Acyclovir (internal standard, IS)
Log P -0.81 -1.66 -0.95
pKa 10.16 8.71 11.98
Reference [4] [5] [6]
Table 1: Log P and pKa for valganciclovir, ganciclovir and acyclovir

Previous analytical methods using HPLC with UV detection [7, 8, 9] and HPLC coupled with mass spectrometer [10] were not able to measure valganciclovir and ganciclovir simultaneously, rather only one substance at a time.

Optimal chromatographic separation with MS detection

The authors of this publication tried to develop a method for simultaneous determination of GAN and VAL using HPLC-UV, but this method required the use of a specific “YMC-Pack Polyamine II” chromatographic column and setting of eluent flow at 2 mL/min, which significantly increased reagent consumption. Moreover, analysis time was quite long (about 26 minutes) to achieve optimal separation of the substances and internal standard.

The combination of MS detection and chromatographic separation enables reducing the flow rate and significantly reducing the analysis time. This study provides development and validation of the methods for the determination of valganciclovir and ganciclovir in human blood plasma by LC-MS/MS.

Materials and methods

The UHPLC Nexera XR coupled with tandem mass spectrometer LCMS-8040 (both from Shimadzu) were used. Methanol (UHPLC-grade) was purchased from J.T.Backer, acetonitrile (LCMS-grade) from Biosolve, formic acid (98 % pure) and aqueous ammonia (“for analysis” grade) from PanReac. Deionized water was produced by “Milli-Q” system from Millipore. Valganciclovir hydrochloride (USP reference standard, 99.2 %), ganciclovir (USP reference standard, 97.5 %) and acyclovir (ACI) applied as an internal standard (USP reference standard, 94.6 %) were used to prepare stock solutions by dissolving in methanol. Mixed working solutions of GAN and VAL and an ACI working solution were prepared by diluting of stock solutions with the same solvent to the required plasma concentrations (table 2). Stock and working solutions as well as intact blood plasma samples were stored prior to use in a freezer at -45 oC.

Level Analyte concentration, ng/mL IS concentration, ng/mL
VAL GAL ACI
1 5.00 50.00 1,000.00
2 10.00 100.00 1,000.00
3 25.00 250.00 1,000.00
4 50.00 500.00 1,000.00
5 100.00 1,000.00 1,000.00
6 250.00 2,500.00 1,000.00
7 400.00 4,000.00 1,000.00
8 750.00 7,500.00 1,000.00
9 1,000.00 1,0000.00 1,000.00
LLOQ 5.00 50.00 1,000.00
L 15.00 150.00 1,000.00
M 500.00 5,000.00 1,000.00
H 800.00 8,000.00 1,000.00
Table 2: Concentrations of target substances at calibration levels

Sample preparation

10 μL of acyclovir working solution was added to 200 μL of a calibration sample placed in 2 mL Eppendorf centrifuge tubes, 600 μL of acetonitrile were then added, vortexed for ten seconds and centrifuged for 15 min at 13,500 rpm. The resulting supernatant was transferred into vials and placed in the autosampler.

HPLC conditions

Column:Shim-Pack GWS C18 (5 µm; 150 x 4.6 mm)
Column temperature:40 °C
Mobile phase (A):0.1 % formic acid in water with 0.08 % ammonia (v/v)
Mobile phase (B):0.1 % formic acid, 10 % water in acetonitrile with 0.08 % ammonia (v/v)
Gradient:The gradient is shown in table 3
Injection volume:5 μl
Time, min. Mobile phase A, % Mobile phase B, % Flow rate, ml/min
0.00 85.00 15.00 1.00
0.70 85.00 15.00
3.50 90.00 10.00
4.00 0.00 100.00
5.50 0.00 100.00
5.70 85.00 15.00
7.00 85.00 15.00
Table 3: Gradient elution

MS conditions

Ionization:ESI, positive
Nebulizing gas:3 L/min
Drying gas:20 L/min
Heater block:400 oC
Desolvation line:200 oC
VAL:355.20 m/z → 152.05 m/z
GAL:255.80 m/z → 152.00 m/z; 255.80 m/z → 134.90 m/z
ACI:226.00 m/z → 152.00 m/z; 226.00 m/z → 134.95 m/z
Registration time:0.00 – 7.00 min

Results and discussion

Method development

Development of a chromatographic separation method for valganciclovir and ganciclovir is challenging due to their weak retention on “classic” octadecyl columns. Use of the columns with amino groups also did not provide good retention and proper separation of VAL or GAL. An idea was to test C18 columns with different total carbon content (table 4). The best results were achieved using Shim-Pack GWS column with the lowest total carbon content. Despite the insignificant retention of valganciclovir and ganciclovir on this column, the chromatographic peaks were separated from the dead volume, and the method met the requirements of the regulation for the validation, including the “selectivity” parameter.

Method validation

The method was validated based on the guidelines for the examination of medicines [12] as well as FDA [13] and EMA [14] guidelines for the following parameters: selectivity, matrix effect, calibration curve, accuracy, precision, recovery, lower limit of quantification, carry-over and stability.

Selectivity

Six samples of intact plasma obtained from different sources were tested as well as samples of intact plasma spiked with working solutions of valganciclovir and ganciclovir (5 and 50 ng/mL respectively). Additionally, samples of intact plasma with hemolysis and samples with increased lipid content were analyzed. With blank plasma chromatograms (figure 3) intensity of the peaks with the same retention time as the target substances did not exceed 20 % of the signal at LLOQ (Lower Limit Of Quantification) and 5 % of the internal standard signal.

Figure 3: Blank plasma sample chromatogram
Сhromatographic column Total carbon content, % Surface area,
m²/g
Phenomenex Luna C18(2)
50 × 2 mm, 5 µm
18,2 393
Phenomenex Luna NH2 50 × 3 mm. 5 µm 10.2 420
Waters XBridge C18
50 × 4.6 mm, 3.5 µm
18.0 178
YMC Hydrosphere C18
100 × 2 mm, 3 µm
12.2 330
Shim-pack GWS
150 × 4.6 mm, 5 µm
9.5 450
Table 4: Some properties of chromatographic columns used in the study

Matrix effect

The samples including working solutions of valganciclovir, ganciclovir and acyclovir as well as spiked plasma samples were examined to estimate a matrix effect. It was estimated at low (L) and high (H) concentration levels (see table 2) for both valganciclovir and ganciclovir and at concentration of 1,000 ng/mL for acyclovir. The results are shown in tables 5 and 6.

# Mf of GAL (Level L) Mf of ACI (Level L) Normalised Mf (Level L) Mf of GAL (Level H) Mf of ACI (Level H) Normalised Mf (Level H)
1 0.91 1.00 0.90 1.00 0.99 1.01
2 0.87 1.00 0.87 0.99 1.02 0.97
3 0.91 1.01 0.90 0.98 0.99 0.99
4 0.92 1.02 0.91 1.01 1.01 1.00
5 0.90 0.99 0.91 0.99 1.02 0.97
6 0.94 1.01 0.93 1.01 1.02 0.99
Average 0.90 Average 0.99
CV, % 2.18 CV, % 1.50
Table 5: The matrix factor of valganciclovir calculations, normalized by the IS matrix factor
# Mf of GAL (Level L) Mf of ACI (Level L) Normalised Mf (Level L) Mf of GAL (Level H) Mf of ACI (Level H) Normalised Mf (Level H)
1 1.18 1.00 1.17 1.04 0.99 1.04
2 0.98 1.00 0.98 0.82 1.02 0.81
3 0.97 1.01 0.97 0.84 0.99 0.85
4 1.14 1.02 1.12 0.99 1.01 0.98
5 1.29 0.99 1.30 0.85 1.02 0.84
6 1.21 1.01 1.19 1.00 1.02 0.98
Average 1.12 Average 0.92
CV, % 11.56 CV, % 10.58
Table 6: The matrix factor of ganciclovir calculations, normalized by the IS matrix factor

Calibration curves

Based on nine samples of intact plasma spiked with acyclovir, valganciclovir and ganciclovir to the concentration levels 1-9 (see table 2) a calibration was created. The calibration curves for valganciclovir (peak area ratio to concentration ratio) and ganciclovir (same coordinates) are shown in figures 5 and 6 respectively. The correlation coefficients obtained meet the regulations (> 0.99). A chromatogram of the sample with concentration level 9 is shown in figure 4.

Figure 5: The calibration curve representing dependence of the ratio area peak of valganciclovir to acyclovir on the concentration ratio of valganciclovir to the acyclovir in plasma
Figure 6: The calibration curve representing dependence of the ratio area peak of ganciclovir to the acyclovir on the concentration ratio of ganciclovir to the acyclovir in plasma
Figure 4: Level 9 plasma sample chromatogram
Figure 7: LLOQ plasma sample chromatogram

Accuracy and precision

Spiked plasma samples at LLOQ, high, medium and low concentration levels (see table 2) were analyzed to estimate accuracy and precision of the method developed. Five different samples for each concentration level with triple injections were used. Accuracy and precision have been evaluated within-run, and between-run (two runs and three runs); data is shown in table 7. The obtained values of the relative standard deviation (precision) and the relative error (accuracy) meet the regulations (no more than 20 % at the LLOQ level, no more than 15 % at other levels).

Injected (ng/mL) Average found, ng/mL SD RSD, % E, %
(n = 5) (n = 10) (n = 15) (n = 5) (n = 10) (n = 15) (n = 5) (n = 10) (n = 15) (n = 5) (n = 10) (n = 15)
Valganciclovir
5.00 5.80 5.45 5.29 0.08 0.62 0.56 1.37 11.35 10.53 16.08 8.96 5.89
15.00 14.46 14.08 13.72 0.62 0.75 0.80 4.26 5.31 5.86 -3.63 -6.11 -8.52
500.00 550.73 561.04 563.16 5.21 12.53 10.69 0.95 2.23 1.90 10.15 12.21 12.63
800.00 893.66 901.00 900.88 12.41 13.06 12.59 1.39 1.45 1.40 11.71 12.62 12.61
Ganciclovir
50.00 48.32 47.39 47.18 1.13 2.33 2.06 2.33 4.91 4.37 -3.35 -5.21 -5.65
150.00 145.73 143.35 141.91 6.62 6.30 5.73 4.27 4.39 4.04 -2.85 -4.44 -5.39
5000.00 4740.25 4808.83 4892.22 77.26 104.66 153.94 1.63 2.18 3.15 -5.19 -3.82 -2.16
8000.00 7198.32 7372.03 7482.18 146.29 225.72 252.41 2.03 3.06 3.37 -10.02 -7.85 -6.47
Table 7: Accuracy and precision of the method

Recovery

Three plasma samples spiked at low, medium and high concentration levels as well as QC samples were tested to estimate recovery value. Additionally, the plasma samples with hemolysis and increased lipid content were tested. The results obtained are presented in table 8. According to the regulations, a recovery should not be 100 %, however, it is necessary to ensure efficient and reproducible extraction of the target substances from the biological matrices. RDS of recovery value should not exceed 15 %.

Recovery (Level L), % Recovery (Level M), % Recovery (Level H), %
Valganciclovir
Average 88.57 98.94 101.80
SD 6.78 2.76 2.48
RSD 7.66 2.79 2.43
Ganciclovir
Average 93.98 90.95 98.01
SD 8.64 9.83 4.44
RSD 9.20 10.80 4.53
Table 8: Calculation of valganciclovir and ganciclovir recovery at L, M, H levels from different biological matrix

Lower limit of quantification
LLOQ level was determined based on linearity, accuracy and precision data. The minimum valganciclovir and ganciclovir concentrations in plasma for which it is possible to quantify VAL and GAL with RSD and E values of no more than 20 % were defined as LLOQ for the method. LLOQ was 5 ng/mL for valganciclovir and 50 ng/mL for ganciclovir. Chromatograms of plasma containing VAL and GAL at LLOQ level are shown in figure 7. The detection limit of valganciclovir was about 0.93 ng/mL, and the detection limit of ganciclovir was about 0.73 ng/mL (signal/noise ratio about 3:1).

Stability

Short-term stability of the samples prepared (autosampler stability and bench-top stability) was confirmed at lower and upper concentration levels. Also, a solution stability of the target substances at three freeze-thaw cycles as well as long-term stability (when stored for 30 and 59 days at a temperature of -45 °C) was confirmed.

Carry-over

There were no peaks with retention time of the target substances on chromatogram when analyzing the blank plasma samples after the calibration samples with the highest concentration level of valganciclovir and ganciclovir.

Conclusion

The LC-MS/MS method for the determination of valganciclovir and ganciclovir in human plasma was developed and validated. Concentration ranges in plasma were 5-1,000 ng/mL for valganciclovir and 50-10,000 ng/mL for ganciclovir. This enables use of the method for both pharmacokinetics studies and therapeutic drug monitoring.

Literature

  1. Karpov I. A., Salavei M. U.
    Valganciclovir as a highly effective drug for the prevention and treatment of herpes infections.
    Meditsinskie novosti. 2018;3:56-59. (In Belarus)
  2. Xu H.-R., Li X.-N., Chen W.-L., Liu G.-Y., Chu N.-N., Yu C.
    A sensitive assay for simultaneous determination of plasma concentrations of valganciclovir and its active metabolite ganciclovir by LC/MS/MS.
    Journal of Chromatography B. 2007;848(2):329-334. DOI: 10.1016/j.jchromb.2006.10.053
  3. Singh O., Saxena S., Mishra S., Khuroo A., Monif T. Determination of valganciclovir and ganciclovir in human plasma by li quid chromatography tandem mass spectrometric detection.
    Clinical Biochemistry. 2011;44(10-11):907-915. DOI: 10.1016/j. clinbiochem.2011.04.003
  4. Valganciclovir. DRUGBANK. Available at: https://go.drugbank.com/drugs/DB01610
  5. Ganciclovir. DRUGBANK. Available at: https://go.drugbank.com/drugs/DB01004
  6. Acyclovir. DRUGBANK. Available at: https://go.drugbank.com/drugs/DB00787
  7. Teshima D., Otsubo K., Yoshida T., Itoh Y., Oishi R.
    A simple and simultaneous determination of acyclovir and ganciclovir in human plasma by high‐performance liquid chromatography.
    Biomedical Chromatography. 2003;17(8):500-503. DOI: 10.1002/bmc.258
  8. Padullés A., Colom H., Armendariz Y., Cerezo G., Caldes A., Pou L., Torras J., Grinyó J. M., Lloberas N.
    Determination of ganciclovir in human plasma by ultra performance liquid chromatography – UV detection.
    Clinical Biochemistry. 2012;45(4-5):309-314. DOI:10.1016/j.clinbiochem.2011.12.014
  9. Heinig K., Wirz T., Gajate-Perez A., Belli S. Determination of Ganciclovir and its prodrug Valganciclovir by hydrophilic interaction liquid chromatography-tandem mass spectrometry. Journal of Chromatography B. 2011;879(5-6):436-442. DOI: 10.1016/j. jchromb.2010.12.033
  10. Dogan-Topal B., Ozkan S. A., Uslu B.
    Simultaneous determination of abacavir, efavirenz and valganciclovir in human serum samples by isocratic HPLC-DAD detection.
    Chromatographia. 2007;66(1):25-30. DOI: 10.1365/s10337-007-0302-y
  11. Komarov T. N., Shohin I. E., Miskiv O. A., Bogdanova D. S., Aleshina A. V., Medvedev Yu. V., Bagaeva N. S.
    Development and Validation of Valganciclovir and its Active Metabolite Ganciclovir Determination in Human Plasma by HPLC-UV Method. Razrabotka i registratsiya lekarstvennykh sredstv = Drug development & registration. 2020;9(2):133-139. (In Russ.) DOI: 10.33380/2305-2066-2020-9-2-133-139
  12. Mironov A. N. Rukovodstvo po ekspertize lekarstvennykh sredstv [Guidelines for the examination of medicines]. V. II. Moscow: Grif i K; 2013: 280 p. (In Russ.)
  13. Food and Drug Administration. Available at: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/bioanalytical-method-validation-guidance-industry
  14. European Medicines Agency. Available at: https://www.ema.europa.eu/en/bioanalytical-method-validation