Paired comparisons of venetoclax concentration in cerebrospinal fluid, bone marrow, and plasma in acute leukemia patients

Abstract

Venetoclax, a small molecule inhibitor of BCL‐2, has demonstrated efficacy in treating acute leukemias and has been recommended as one of the first‐line anti‐leukemia therapies. Although venetoclax has been suggested to probably possess the ability to penetrate the central nervous system (CNS), current data to elucidate the characteristics of venetoclax in cerebrospinal fluid (CSF), bone marrow (BM), and plasma are still lacking. This study investigated the real‐world characteristics of venetoclax concentrations in CSF, BM, and plasma in acute leukemia patients. Thirteen acute leukemia patients treated with venetoclax were included, with paired samples of CSF, BM, and plasma collected and venetoclax concentrations measured using LC–MS/MS. With the results, the median venetoclax concentrations were 2030 ng/mL in plasma, 16.7 ng/mL in CSF, and 1390 ng/mL in BM. The percentages of CSF/plasma and BM/plasma were 0.74% and 70.37%, respectively. While no direct correlation was observed between CSF and plasma venetoclax levels, there was a trend toward an improved CSF/plasma percentage over time following the last administration of venetoclax. In contrast, a strong correlation was found between BM and plasma levels. This study demonstrated that venetoclax could reach its effective concentration in most patients, suggesting its potential clinical utility in the management of CNS involvement in acute leukemia.

Study Highlights.

• WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?

Venetoclax has demonstrated efficacy in treating acute leukemias and has been recommended as one of the first‐line anti‐leukemia therapies. Although venetoclax has been suggested to probably possess the ability to penetrate the central nervous system (CNS), current data to elucidate the characteristics of venetoclax in cerebrospinal fluid (CSF), bone marrow (BM), and plasma are still lacking.

• WHAT QUESTION DID THIS STUDY ADDRESS?

This study investigated the real‐world characteristics of venetoclax concentrations in CSF, BM, and plasma in acute leukemia patients.

• WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?

This study demonstrated venetoclax could reach its effective concentration in most acute leukemia patients, suggesting its potential clinical utility in the management of CNS involvement in acute leukemia.

• HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?

For novel anti‐leukemia agents, monitoring their concentrations at different sites is crucial for guiding the selection of therapeutic agents for various clinical disease types.

INTRODUCTION

Central nervous system (CNS) involvement in acute leukemia is a high‐risk clinical course with limited treatment options and a dismal prognosis. ¹ , ² Current therapy for CNS leukemia involves intrathecal chemotherapy, intensive systemic chemotherapy, and cranial radiotherapy. ³ Nevertheless, a certain proportion of patients will become resistant to these therapies, indicating that new effective treatment options for CNS leukemia are urgently needed. Venetoclax is a small molecule that selectively inhibits the antiapoptotic protein BCL‐2, ⁴ which has been recommended as the first‐line anti‐leukemia therapy ⁵ and has demonstrated significant antileukemic activity in both acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL). ⁶ , ⁷ , ⁸ If venetoclax can be proved to penetrate the blood–brain barrier, it would be of great value in the treatment of CNS leukemias. However, current evidences of venetoclax in cerebrospinal fluid (CSF) were based on a phase I study or case reports. ⁹ , ¹⁰ , ¹¹ , ¹² Data to elucidate the characteristics of venetoclax in CSF, bone marrow (BM), and plasma are still lacking. Therefore, we conducted a study aimed to explore the real‐world characteristics of paired venetoclax concentration in CSF, BM, and plasma in acute leukemia patients.

METHODS

In this study, paired samples of CSF, bone marrow, and plasma in acute leukemia patients treated with venetoclax were collected at the same time. LC–MS/MS method was used to detect venetoclax concentration both in plasma and CSF. LC‐20AXR (Shimadzu, Japan) tandem AB SCIEX API 5500 (AB Applied Biosystems, USA) was used in this experiment. The data were acquired and processed using Analyst software (AB SCIEX, version 1.5). The retention times of venetoclax and ²H₇‐venetoclax were 3.0–3.2 min. The positive electrospray ionization (ESI+) mode was selected in our method with the fragments of m/z = 868.5/321.0 for venetoclax, and m/z = 875.5/321.0 for IS. The calibration curve of venetoclax was validated at (30, 60, 100, 200, 500, 600, 1000, 3000, 8000) nine points over the concentration range of 30–8000 ng/mL. The linear regression equation (“1/(x*x)” weighing) is y = 0.0073x + 0.00643 (r ² = 0.9980).

Statistical analysis was performed using Graphpad Prism 6 software (San Diego, CA, USA). The correlation between venetoclax concentrations in plasma, CSF, and BM, and the correlation between CSF/plasma percentage and time were analyzed by the linear regression model. Statistical significance was considered if p‐value was less than 0.05. The protocol had been approved by the Beijing Chaoyang Hospital Ethics Committee. All patients had written informed consent before treatment.

RESULTS

From October 16, 2023, through January 22, 2024, a total of 13 patients from Beijing Chaoyang Hospital were enrolled in this study. Baseline characteristics are listed in Table 1. Eight patients (61.5%) were male, and five (38.5%) were female, with a median age of 47 years (range: 15–75). All of the 13 patients were diagnosed with acute leukemia. Among which these 10 patients (76.9%) were AML, including 5 (38.5%) acute promyelocytic leukemias (APL) and 5 (38.5%) non‐APL AMLs while the other 3 patients (23.1%) were ALL. Paired venetoclax concentrations were detected twice in two patients (Patient 1 and 5) in different treatment cycles. Among all of the patients, venetoclax concentrations were detected after a median of 7 days (range: 4–14) from the start of continuous venetoclax administration, and a median of 20 h (range: 4–48) since the last dose of venetoclax. In 92.3% (12/13) of the patients, venetoclax was administrated with a standard dose of 400 mg/day. While in one patient (Patient 3), venetoclax dosage was 100 mg/day combined with voriconazole 400 mg/day. Combination therapy included dasatinib, ponatinib, prednisone, homoharringtonine, cytarabine, aclarubicin, azacitidine, doxorubicin, ivosidenib, and granulocyte colony‐stimulating factor (G‐CSF).

TABLE 1.

Clinical characteristics and venetoclax concentrations.

No.	Sex	Age	Diagnosis	Days of VEN	Dose of VEN (mg)	Time since last VEN dose (h)	Combination therapy	Combination of azoles	VEN in Plasma (ng/mL)	VEN in CSF (ng/mL)	VEN in BM (ng/mL)	CSF/Plasma	BM/Plasma
Patient 1 (1st)	Male	22	ALL	14	400	48	Dasatinib, prednisone	N/A	128	13.4	77.8	10.47%	60.78%
Patient 1 (2nd)	Male	22	ALL	14	400	24	Dasatinib, prednisone	N/A	2970	22.0	2090	0.74%	70.37%
Patient 2	Male	15	APL	7	400	26	Homoharringtonine, cytarabine, G‐CSF	N/A	2270	18.7	1390	0.82%	61.23%
Patient 3	Male	47	APL	6	100	23	Aclarubicin, cytarabine, G‐CSF	Voriconazole 400 mg/d	2670	25.3	–	0.95%
Patient 4	Male	30	APL	6	400	6	Aclarubicin, cytarabine, G‐CSF	N/A	1890	12.5	–	0.66%
Patient 5 (1st)	Female	75	AML	7	400	5	Azacitidine, G‐CSF	N/A	2030	22.8	–	1.12%
Patient 5 (2nd)	Female	75	AML	8	400	20	Azacitidine, G‐CSF	N/A	2580	9.1	–	0.35%
Patient 6	Female	36	AML	7	400	24	Doxorubicin, cytarabine	N/A	524	24.7	–	4.71%
Patient 7	Male	61	AML	5	400	22	Azacitidine, G‐CSF	N/A	844	22.7	–	2.69%
Patient 8	Female	65	AML	4	400	16	Azacitidine, G‐CSF	N/A	5530	26.5	–	0.48%
Patient 9	Male	58	ALL	7	400	4	Ponatinib, prednisone	N/A	4270	16.4	3250	0.38%	76.11%
Patient 10	Male	46	APL	7	400	22	Aclarubicin, cytarabine, G‐CSF	N/A	528	16.7	–	3.16%
Patient 11	Female	41	APL	7	400	13	Aclarubicin, cytarabine, G‐CSF	N/A	1670	10.0	–	0.60%
Patient 12	Male	55	ALL	7	400	14	Dasatinib, prednisone	N/A	1760	5.3	1290	0.30%	73.30%
Patient 13	Female	66	AML	5	400	20	Ivosidenib, azacitidine, G‐CSF	N/A	2790	4.4	–	0.16%

Abbreviations: ALL, acute lymphocytic leukemia; AML, acute myelocytic leukemia (non‐APL); APL, acute promyelocytic leukemia; BM, bone marrow; CSF, cerebrospinal fluid; G‐CSF, granulocyte colony‐stimulating factor; VEN, venetoclax.

In all 15 plasma‐CSF paired samples, the median venetoclax plasma concentration was 2030 ng/mL (range: 128–5530), while median venetoclax CSF concentration was 16.7 ng/mL (range: 4.4–26.5). The median CSF/plasma venetoclax percentage was 0.74% (range: 0.16%–10.47%). There were also five BM samples which had been detected with venetoclax concentration along with CSF and plasma. The median venetoclax BM concentration was 1390 ng/mL (range: 77.8–3250), while the median BM/plasma percentage was 70.37% (range: 60.78%–76.11%) (Table 1). The correlations of concentrations between plasma, CSF, and BM are shown in Figure 1c. Although venetoclax concentration between CSF and plasma did not show a significant linear correlation (R ² = 0.027, p = 0.558) (Figure 1b), venetoclax concentration between BM and plasma showed a significant linear correlation (R ² = 0.985, p < 0.001) (Figure 1a). However, there was a trend toward an improved CSF/plasma percentage over time following the last administration of venetoclax (R ² = 0.601, p < 0.001) (Figure 1d). Notably, in patients 1 and 5, paired samples of plasma and CSF had been detected twice in two different venetoclax cycles. The percentages of CSF/plasma showed discrepancy at different timepoints within the same patients (10.47% vs. 0.74%, and 1.12% vs. 0.35%), indicating that venetoclax in CSF did not directly correlate with its plasma concentration.

FIGURE 1.

Venetoclax concentration in plasma, CSF, and BM in acute leukemia patients. (a) Correlation between plasma and BM venetoclax concentrations. (b) Correlation between plasma and CSF venetoclax concentrations. (c) Correlation between plasma, CSF, and BM venetoclax concentrations. (d) Correlation between venetoclax CSF/plasma percentage and time since last venetoclax dose.

DISCUSSION

To the best of our knowledge, this is the first study to describe the real‐world characteristics of venetoclax CSF and BM concentration in adult leukemia patients. In our study, the median venetoclax CSF concentration was 16.7 ng/mL, and the median percentage of CSF/plasma venetoclax was 0.74%, which was higher than the results of a phase I study reported by Badawi et al. ¹² in pediatric patients, with a median CSF concentration of 1.5 ng/mL, and a median plasma/CSF ratio of 285 (CSF/Plasma: 0.35%). Possible explanations for the discrepancy might be the differences in patients’ age (47 vs. 8.5 years) and venetoclax duration. In our study, venetoclax concentration was detected in a median of 7 days after venetoclax continuous administration. While in the study of Badawi et al., ¹² although median venetoclax duration was not specified, pharmacokinetic samples were mainly collected on days 1–3 and day 8 per the protocol, ¹³ which might result in the differences in venetoclax CSF concentration.

Based on current evidences, venetoclax exhibits the capacity to penetrate into blood–brain barrier. ⁹ , ¹¹ , ¹² Taking into consideration that venetoclax is highly protein bound with less than 1% free unbound drug, there was limited free drug that is available to distribute into tissues, our data that venetoclax in CSF accounted for about 0.7% of its plasma concentration showed a relatively high penetration ability of venetoclax into CNS. According to the literature, venetoclax half‐maximal inhibitory concentrations (IC50) vary across subtypes of acute leukemia. ⁷ , ¹⁴ , ¹⁵ Venetoclax concentration in CSF (median 16.7 ng/mL, i.e., ~19.2 nmol/L) detected in our study had reached its median IC50 in vitro in about half of AML and ALL cell lines, ⁷ , ¹⁴ and for primary patient AML myeloblasts. ¹⁵ Moreover, we observed a trend toward an improved CSF/plasma percentage over time following the last administration of venetoclax, indicating that venetoclax might have a longer half‐life in CSF than in plasma, which demonstrated the potential anti‐leukemia efficacy of venetoclax in CNS leukemia.

In some subtypes of acute leukemia such as AML with RAS mutation, which had high IC50 levels and was associated with venetoclax resistance, the effect of venetoclax in CNS leukemia might be attenuated.

Furthermore, venetoclax concentrations in BM were detected in five paired samples and showed a comparable level in plasma and BM, with a median BM/plasma percentage of 70.37%. Since bone marrow is where the blasts are produced and is critical for leukemia occurrence, the high penetration of venetoclax in BM could explain its efficacy in acute leukemias. ⁷ , ⁸

In summary, our study analyzed venetoclax concentration in paired samples from CSF, BM, and plasma in acute leukemia patients and demonstrated venetoclax could reach its effective concentration in most patients, suggesting its potential clinical utility in the management of CNS involvement in acute leukemia.

CONFLICT OF INTEREST STATEMENT

The authors declared no competing interests for this work.

Jian Y, Han F, Zhu Y, et al. Paired comparisons of venetoclax concentration in cerebrospinal fluid, bone marrow, and plasma in acute leukemia patients. Clin Transl Sci. 2024;17:e70006. doi: 10.1111/cts.70006

DATA AVAILABILITY STATEMENT

All data generated or analyzed during this study are included in this published article.