Supplemental Pioglitazone to Patients of CML with Suboptimal TKI Response: A Pragmatic Pilot Study

Abstract

Tyrosine kinase inhibitors (TKIs) have improved outcomes of chronic myeloid leukemia (CML). However, 20–30% of patients require second-line TKIs following suboptimal response. The cost and adverse events limit their use in resource-constraint settings. We conducted a pilot study to ascertain the benefit of adding pioglitazone to TKIs with suboptimal response in real-world resource-constraint settings. In this pragmatic pilot study from 01 Jan 2017 to 31 July 2021, CML patients from a tertiary care center in North India with sub-optimal response to TKIs were additionally given pioglitazone after ruling out imatinib resistance mutation (n − 31). Pioglitazone was stopped if there was disease progression on follow-up, and second-line TKI was started. The data were analyzed with the intention-to-treat principle using JMP Ver.15.1.1. The median age of the study population was 54y (27–82), who were followed up for a median duration of 1023.5d (59–1117). Pioglitazone showed the benefit of one-log reduction in BCR-ABL in 89.7% of the study participants. 1y, 2y and 3y-PFS were 92.57%, 76.5%, and 68.3% respectively. During follow-up period, the disease progressed in 38.7%, of which two succumbed. No adverse events to Pioglitazone were documented. This study proved that adding Pioglitazone to the existing TKI regime in CML with sub-optimal response can benefit. The addition of Pioglitazone was well tolerated.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s12288-022-01561-x.

Introduction

Chronic Myeloid Leukemia (CML) is a chronic myeloproliferative neoplasm with hallmark Philadelphia chromosome with translocation t(9;22) (q34;q11.2). The resulting Breakpoint Cluster Region-Abelson (BCR-ABL1) fusion oncoprotein, with its constitutive tyrosine kinase activity, leads to activation of downstream signaling pathways like Rat Sarcoma Virus (RAS), Rapidly Accelerated fibrosarcoma (RAF), Myelocytomatosis (MYC), Signal Transducer and Activator of Transcription (STAT). The Tyrosine kinase inhibitors (TKIs) were developed in the late 90s and changed the whole treatment landscape, improving the 10-year survival rate to at par with gender/age-matched population without CML [1–3]. However, CML is incurable, requiring life-long TKI therapy, and 20–30% of patients do not achieve optimal response with imatinib [4]. With the advent of generic imatinib, the therapy became affordable in resource-poor settings. The cost of therapy significantly rises with second-line therapies, which at times is not financially feasible in developing countries like India.

Pioglitazone, a peroxisome proliferator-activated receptor-gamma agonist (PPAR-γ), was implicated in achieving complete molecular response (CMR) in three CML patients who had not achieved CMR even after 4–6 years of Imatinib therapy in 2015 [5]. In another study of 24 CML-CP patients, the authors demonstrated the achievement of deep molecular response (DMR; defined as ≥ MR^4.5) by 12 months in 54% of patients using pioglitazone [6]. Though these studies highlight the benefit of pioglitazone in patients with an optimal molecular response for achieving a deeper response, there is presently a lack of data on the efficacy of adding pioglitazone to suboptimal molecular response to an optimized TKI regime [7].

Thus, we aimed to study the effect of adding pioglitazone to TKIs in CML patients who failed to achieve optimal response at any time point and observe its safety/efficacy.

Materials and Methods

This prospective, open-label, interventional pilot study was carried out at a tertiary care center in North India from 01 Jan 2017 to 31 July 2021. The study was approved by the institutional ethics committee and was conducted in accordance with the Declaration of Helsinki. All patients of CML from either gender, aged more than 18 years, with no significant liver function derangements and a suboptimal molecular response, on the optimal dose of TKI (either Imatinib, Nilotinib, or Dasatinib) were screened for eligibility. The suboptimal molecular response was defined as not achieving BCR/ABL(IS) < 10% at 03 months or < 1% at 06 months or < 0.1% (MR³) at any time after 12 months [i.e., values in a warning and failure range to TKI as per European Leukemia Net (ELN) guidelines] or 01-log rise in BCR/ABL(IS) while on therapy [8]. Before inclusion, the study population was subjected to Imatinib Resistance Mutation Analysis (IRMA). IRMA or kinase domain mutation analysis was done by sanger sequencing and was outsourced due to a lack of an in-house testing facility. Those with specific mutations were shifted to specific second-line TKIs based on the sensitivity (standard of care). Those negative for mutations on IRMA only were included in the study. In this study, those patients with IRMA positivity were excluded, as the IEC had disapproved the deviation from the standard of care for investigating an experimental therapy.

All patients were also subjected to baseline ECG, 2D Echo, CBC, LFT, Blood sugar fasting, and postprandial and fibroscan. Those patients enrolled in any other investigational drug trial were excluded from the study. At enrolment, any patients with previous or newly diagnosed diabetes mellitus, cardiovascular disease in the form of congestive heart failure, myocardial infarction within six months of enrolment, cardiac arrhythmias, liver disease (deranged LFT more than 2.5 times above upper lab limit or Fibroscan > 9.5 kPa), history bladder cancer or hematuria, macular edema or uncontrolled peripheral edema of any etiology were excluded from the study. Tab Pioglitazone 30 mg/day was added to their existing TKI regime. These patients were followed at 3, 6, 12, and 18 months; after that six-monthly review with repeat quantitative BCR-ABL with international standardization [RQ-PCR BCR-ABL (IS)], complete blood count, biochemical parameters including liver functions, and fibroscan. If there was disease progression as per BCR-ABL as per existing guidelines [8] transcript values, the Pioglitazone was stopped, and a second line TKI was started as the standard of care. Progression was defined as loss of complete molecular response or even a rise of more than one log on follow-up BCR-ABL RQPCR was taken as disease progression. The patients were monitored for the hypoglycemia and hepatotoxicity more stringently, wherein during the initial four weeks, patients were followed up weekly, after that, the patients were followed up monthly for three months and later quarterly. The patients were monitored for urinary bladder cancer with quarterly sonography of the urogenital system (with particular attention to the urinary bladder) and diligent history of hematuria during every clinical visit. Time-related adverse events were recorded as per Common Terminology Criteria for Adverse Events (CTCAE) v4.0. Any adverse events ≥ 3 on CTCAE grading were considered clinically significant.

The endpoints studied were the number of patients achieving 1 log reduction of RQ-PCR BCR-ABL (IS), the number of participants with treatment-related adverse events as assessed by CTCAE v4.0, and the assessment of progression-free survival.

The data were analyzed with the intention-to-treat principle using JMP Ver.15.1.0 and R Ver.3.4.0 [9]. All continuous variables were evaluated for distribution, and thereafter, normally distributed variables were analyzed using parametric tests and others using non-parametric tests. Kaplan–Meier curve was used for reporting progression-free survival (PFS). Logistic regression analysis using linear and the lowess fit was used to study the effect of the intervention (pioglitazone) on BCR-ABL over the study period.

Results

We screened 452 patients with CML who had a suboptimal response to their present TKIs, of which 257 patients were found to have underlying IRMA and were shifted to second-line TKI (Fig. 1). Of the rest of the patients, only 31 patients met the inclusion/ exclusion criteria and were enrolled in the study after informed consent.

Fig. 1.

Diagram showing selection of the study participants

The median age of the study population was 54y (27–82), with a male preponderance (74%, n − 23). Of the study group, 16% of patients were on dasatinib, 19% were on nilotinib, and the rest were taking imatinib at the time of enrolment. The patients were enrolled after a median of 1511 days (193–5427) from the initial diagnosis of CML. The study’s median follow-up was 1023.5 days (59–1117).

Pioglitazone had shown benefit (one log reduction in BCR-ABL) in 89.7% of patients (n − 26). The median time to first benefit was 121 days (0–549 days), and the median duration of benefit was 602 days (0–1083 days). A total of 48.3% of patients (n − 15) were in major molecular response, and 19.3% (n − 6) patients were in deep molecular response (≥ MR^4.5) at the time of censoring the data. On regression analysis, both linear fit and LOWESS fit at six-monthly intervals showed significant benefit (Fig. 2) with a steady and gradual fall in BCR-ABL over time.

Fig. 2.

Linear regression with linear fit and LOWESS fit of Log BCR-ABL and time in days

The absolute mean log reduction in BCR-ABL and mean BCR-ABL percentage at six-monthly intervals from enrollment in the study is as shown in Table 1. There was no documented clinically significant adverse event to pioglitazone in the form of hypoglycemia, hematuria, or development of other malignancies (Bladder cancer) during the follow-up period. Of the total study population, 38.7% of patients had disease progression (n − 12), of which two patients succumbed to their illness.

Table 1.

Absolute mean log reduction in BCR-ABL and mean BCR-ABL percentage at six-monthly intervals

	Mean Log BCR-ABL change		Mean BCR-ABL percentage
	Mean log BCR-ABL	95% LCL	Mean BCR-ABL (IN %)	95% LCL
0.5 years	− 0.57	− 0.80 to − 0.34	0.27	0.16–0.45
1.0 years	− 0.80	− 1.00 to − 0.61	0.16	0.10–0.24
1.5 years	− 1.04	− 1.26 to − 0.82	0.09	0.05–0.15
2.0 years	− 1.28	− 1.56 to − 0.99	0.05	0.03–0.10

The 1y, 2y, and 3y-PFS were 92.57%, 76.5%, and 68.3%, respectively, with a median PFS of 937 days (Fig. 3a). The difference between the PFS based on different types of TKIs backbone to which pioglitazone was added was not statistically significant (Fig. 3b).

Fig. 3.

Kaplan–Meier curve depicting progression-free survival of the a complete study population and b patients classified based on baseline TKI regimen to which pioglitazone was added

Discussion

The present recommendation of lifelong TKI in patients of CML and relapse rates of 40–60% observed in treatment-free remission trials are attributed to the presence of quiescent CML (Ph+) stem cells that are inherently resistant to various TKIs [10–12]. Multiple pathways have been implicated in the survival of CML stem cells, including Hedgehog, NOTCH, GATA-2, and Wingless-related integration site (WNT). However, clinical trials of molecules inhibiting these pathways have not yet demonstrated clinically significant outcomes. In 2015, Prost et al. had implicated PPAR-γ agonist pioglitazone in achieving complete molecular response (CMR) in three CML patients with diabetes [5]. The molecular mechanism for the action of Pioglitazone described in CML is that PPAR-γ agonism leads to downregulation of STAT-5 and subsequent downstream downregulation of HIF-2α and CITED. These pathways are upregulated in CML stem cells to keep them in a quiescent phase. By downregulating these pathways, pioglitazone mobilizes CML stem cells to divide actively and hence amenable to the action of TKI. However, as the outcomes were analyzed in this study with historic imatinib-only cohorts, this was put forth as a likely pathway to stimulate the proliferation of quiescent stem cells leading to depletion of CML stem cell reserve [13].

The ACTIM study was a phase 2 proof of concept study had studied 24 CML-CP patients aged 24–79 years who were in MMR defined by BCR-ABL1 IS ≤ 0.1% but had not achieved DMR after two years of stable imatinib therapy. After adding pioglitazone for one year, the study demonstrated a cumulative incidence of DMR to be 56% by 12 months and in 88% of 17 evaluable patients maintaining the DMR at 48 months on imatinib [6]. Our study demonstrated a significant benefit of adding Tab Pioglitazone to an existing TKI regime in patients with a suboptimal response. The patients in our study were on TKIs other than imatinib, unlike the ACTIM study to simulate the real-world scenario. Only 35% of the patients were on nilotinib/ dasatinib, thus prohibiting a subset analysis. Most patients on these newer TKIs (nilotinib/ dasatinib) had not achieved adequate response with previous imatinib therapy. None of the patients in our study were on second-line TKI upfront as per institutional policy. Our patients were continued on pioglitazone for the entire study period, unlike in ACTIM, where it was stopped after one year. The study demonstrated a sustained one-log reduction for a median duration of about 1.6 years (602 days) and a median PFS of about 2.5 years (937 days).

The patient population subset in the ACTIM study was already in MMR and had persistence of DMR (MR^4.5) at 48 months. In contrast, our study had a patient population with a suboptimal response at the initiation of the study; 48.3% of patients (n − 15) were in MMR, and 19.3% (n − 6) patients were in DMR at the time of censoring the data. The mechanism of persistent molecular response even after stopping the medication in the ACTIM study and non-achievement of DMR in our population subset may be due to differences in inherent sensitivity to TKI between the two groups.

A study by Malhotra et al. comprising 12 patients of CML-CP, with eight patients on imatinib and four patients on nilotinib with the suboptimal response, reported a more than 50% decrease in BCR/ABL transcript in 11 patients, more than one-log reduction in four patient and undetectable levels of BCR-ABL in two patients at six months of follow up, after adding pioglitazone 30 mg once a day to the existing TKI [14]. No significant adverse events were reported, with only one patient having transient grade I transaminitis, not requiring dose reduction or stoppage [14]. Another study from India by Goyal et al. in 31 CML-CP patients not achieving MMR after 12–15 months of imatinib monotherapy, after the addition of pioglitazone 30 mg once a day, demonstrated more than 50% reduction in 16 patients, with seven patients achieving MMR after six months of follow up [7]. One-log reduction was documented in 89.7% of the patients in our study. No significant adverse event was documented related to pioglitazone during the study period.

The study’s major limitations are its limited sample size, the inadequate follow-up period for observing malignancies secondary to pioglitazone, lack of control arm (either placebo/standard of care), and inadequate subset comparison between the different types of TKIs. The important strengths of the study are its three years of follow-up and evaluation of new drug in a resource constraint real-world setting.

In the trial setting, Pioglitazone was shown to reduce BCR-ABL in patients with prior optimal response to TKI. This pilot study showed that the addition of pioglitazone steadily reduces the BCR-ABL in patients having a suboptimal response to TKIs and delays progression, necessitating a change to second-line TKI with no significant clinical adverse events. In a resource constraint setting, adding pioglitazone to boost response and delay progression in patients with suboptimal response to TKIs appears to be a financially viable option if no inherent resistance is detected to a particular TKI, however, it requires corroboration from further studies before implementation in regular practice.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors Contribution

YU was involved in the conception of the study, patient management, data entry, digitalization of records, statistical analysis, manuscript preparation, and review. SP, RK, KM, HK, SS, & SD were involved in patient management, manuscript preparation, and review. NY, HK, & RK were involved in data entry, digitalization of records, and manuscript preparation. NY, RK, & SP were involved in statistical analysis and manuscript review. NY, & RK was involved in manuscript preparation and review.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declarations

Conflict of interest

The authors declare that they have no conflict of interest.