The D₁/D₅ dopamine partial agonist PF-06412562 in advanced-stage Parkinson’s disease: a feasibility study

Abstract

Background:

Current drug treatments have limited efficacy and intolerable side effects in advanced-to-end-stage Parkinson’s disease (advPD). D₁ agonists have the potential to provide symptomatic benefit, but there are no data on safety and feasibility in this population.

Methods:

A two-week, randomized, double blind, crossover phase Ib study was performed in advPD patients to compare standard-of-care (SoC) carbidopa/levodopa with PF-06412562, a selective D₁/D₅ dopamine receptor partial agonist. Each week, there was a Day 1 baseline evaluation with overnight levodopa washout, then treatment on Days 2 and 3 with either SoC or PF-06412562 (split dose 25+20 mg), followed by discharge on Day 4. Primary endpoints were safety and tolerability. Secondary endpoints were global clinical impression of change (GCI-C) rated by clinicians and caregivers.

Results:

Eight advPD patients and their caregivers consented to participate and six were randomized (average disease duration: 22 y). None withdrew voluntarily. One participant with baseline Day 1 dehydration, pre-renal kidney injury, and autonomic dysfunction experienced symptomatic and serious hypotension after receiving PF-06412562 in Week 1 and was discontinued from the study. All other adverse events were rated mild (PF-06412562: n=1, SoC: n=0), moderate (PF-06412562: n=1, SoC: n=1), or severe but non-serious (PF-06412562: n=3, SoC: n=2). No clinically meaningful laboratory changes were observed. Among the five participants who completed the study, GCI-C favored PF-06412562 in two per clinicians’ and four participants per caregivers’ rating.

Conclusions:

PF-06412562 was tolerated in advPD patients. This study provides the feasibility for future safety and efficacy studies in this population with unmet needs.

Trial Registration#:

ClinicalTrials.gov:NCT03665454

Introduction

The dopamine precursor levodopa has provided marked symptomatic relief and prolonged life expectancy in Parkinson’s disease (PD) [1], but with progressive loss of dopaminergic neurons, conversion of levodopa to dopamine becomes less efficient/predictable. Coupled with “off-target” effects [2], there is loss of responsiveness and/or inability to tolerate levodopa. Newer levodopa formulations are somewhat helpful, but do not address the fundamental challenges and often render patients wheelchair-/bed-bound [3, 4]. Theoretically, this could be overcome by mimicking dopamine neurotransmission directly, using a class of drug commonly called “dopamine agonists” to activate postsynaptic dopamine receptors. Indeed, there are many “dopamine agonists” (e.g., pramipexole, ropinirole, etc.) that are or have been approved. In actuality, all approved “dopamine agonists” are misnamed because they are more selective for the dopamine D₂-like subfamily (especially the D₂ and D₃) [5], but not D₁-like (D₁ and D₅) subfamily. Although they reduce motor fluctuations and levodopa-induced dyskinesia in early PD, they are much less efficacious than levodopa [6], and cause intolerable side effects such as hallucinations and confusion that are linked to D₂-like receptor stimulation in a significant fraction of advanced-to-end stage PD (advPD) [7–9].

The distribution of dopamine receptors in the basal ganglia is known to be highly segregated. Unlike D2-receptors, D₁ receptors are expressed almost solely on postsynaptic striatal GABAergic neurons of the direct pathway that we hypothesized were of specific importance in treating PD [10]. Because the striatal GABAergic cells do not degenerate as do nigral dopamine neurons, we hypothesized that D₁-like agonists could be highly efficacious antiparkinson drugs [11, 12]. The first generation of selective full D₁ agonists (dihydrexidine and ABT-431) led to proof-of-concept data by their profound antiparkinsonian effects in both MPTP-treated non-human primates [13] and PD patients [14–16].

Recent discovery of orally available D₁ agonists [17], the second generation of D₁ agonists, has overcome many pharmacokinetic limitations of the first generation compounds [6, 18] and has the potential to impact clinical practice. Promisingly, one of these second generation D₁ agonists is in early Phase III trials. The focus of the ongoing D₁ trials is not advPD, but earlier-stage patients who still respond to levodopa. This choice is understandable because advPD patients are medically complex, and to our knowledge, have never been the subject of a controlled interventional trial despite their extremely poor quality of life and their unmet needs.

The following lines of preclinical evidence suggest that D₁ agonists will be beneficial for advPD patients. First, the D₁ agonist dihydrexidine was more effective in severely parkinsonian MPTP non-human primates than in moderately affected animals [19]. Second, this full D₁ agonist caused profound antiparkinson effects in two MPTP-treated primates unresponsive to levodopa or the D₂ agonist bromocriptine [20]. The availability of efficacious orally available second generation D₁ agonists with excellent pharmacokinetics (e.g., PF-06649751 and PF-06412562 [21]) made it both essential and imperative to determine if studies of D₁ agonists can be conducted in patients with advPD, and to assess whether advanced patients/family were willing and able to participate.

Based on pharmacokinetic considerations, we chose to use PF-06412562 rather than a related compound PF-06649751. Although the latter has somewhat higher D₁ intrinsic activity, it is currently dosed via a tapering over several weeks to achieve full therapeutic levels [22], whereas PF-06412562 showed acute antiparkinsonian effects in 13 PD patients and was well-tolerated at a 50 mg oral dose (t_½ 6.4 h) split into 30 and 20 mg doses given four hours apart, eliminating the need for gradual tapering [23]. To provide proof-of-concept that it is feasible to study D_1/5 agonists in advPD, we conducted a randomized, double blind, crossover, phase Ib, mixed methodology, clinical trial evaluating PF-06412562 versus standard-of-care (SoC) carbidopa/levodopa in advPD patients.

Methods

Trial drug

PF-06412562 (Figure 1) is a non-catechol, D_1/5 partial agonist (K_I 95 nM for human D₁ receptor; adenylate cyclase EC50 580 nM; E_max 44%) The drug and matched placebo were provided by Pfizer. Based on a previous study showing safety and tolerability of a 50 mg oral split dose of PF-06412562, [23], we chose a slightly lower total daily dosage of 45 mg (split 25 and 20 mg) for advPD patients. Sinemet (carbidopa/levodopa, 25/100 mg, encapsulated) was used as reference, with SoC dose based on each participant’s pre-trial total levodopa usage.

Figure 1.

Chemical structure of PF-06412562: (4-[4-(4,6-dimethyl-5-pyrimidinyl)-3-methylphenoxyl]-1H-pyrazolo[4,3-c]pyridine

Trial oversight

This study was conducted at Penn State Health in compliance with the ethical principles of the Declaration of Helsinki and the guidelines for Good Clinical Practice issued by the International Conference on Harmonization. It was reviewed and approved by the US Food and Drug Administration and PSH Institutional Review Board. All participants and/or their caregivers provided signed informed consent. The Data Safety Board (DSB) reviewed safety data after each participant completed the study. The investigators attest to the fidelity of the trial protocol and the authors vouch for the completeness and accuracy of the data. The authors, funder (Pfizer), and current licensee of study drug (Cerevel) agreed to the manuscript submission.

Subjects

Participants were recruited between August 2018-January 2019 from the PSH Movement Disorders Clinic or via a local PD support group. The inclusion and exclusion criteria are detailed in Supplemental Table S1. All met published diagnostic criteria [24], had disease duration >15 years since diagnosis, and were Hoehn & Yahr Stage >IV, even when taking levodopa or when deep brain stimulation was turned on. After consenting, participants and their caregiver were admitted to the Clinical Research Center (CRC) for four days during two consecutive weeks. On Day 1, clinical history, vital signs, blood work [complete blood count (CBC) and chemistry (Na, K, HCO₃, Ca, Mg, AST, ALT, BUN and creatinine)], urinalysis, and a 12-lead electrocardiogram (ECG) were obtained to serve as baseline and assess for significant and/or unstable medical conditions.

If systolic blood pressure (SBP) was <100 mm Hg with evidence of dehydration (BUN/creatinine ratio >20) on Day 1, or SBP was <90 mm Hg with clinical symptoms (dizziness, fainting) on Days 2–4, normal saline was given intravenously until SBP reached >100 mm Hg. Since patients might have been under-medicated during the trial, they were given Sinemet (for parkinsonian symptoms) if patient and/or caregiver deemed that he/she needed it. In addition, acetaminophen (for pain), ondansetron (for nausea), and diphenhydramine (for allergies) were given throughout the study when needed.

Trial design

All study procedures, including a 1-month follow-up phone call, were conducted between August 2018-April 2019. Eligible participants were randomized to either PF-06412562 (Sequence A) or carbidopa/levodopa (Sequence B) in Week 1 (Figure 2) by PSH Investigational Drug Service (IDS) personnel using a 1:1 random allocation sequence (detailed in Supplemental Text S1). Participants, caregivers, and investigators were blinded to sequence assignment.

Figure 2. Schematic of the overall design of the study.

Subjects were randomized to receive PF-06412562 followed by levodopa (Sequence A, top) or levodopa followed by PF-06412562 (Sequence B). PF-06412562 (5 mg) or placebo tablets were provided by Pfizer. Sinemet (carbidopa/levodopa, 25/100 mg) tablets were encapsulated to preserve study blind. The bottom part of the schematic shows the events that occurred on each day. Abbreviations: SoC = Standard of Care. Some subjects received a fourth dose of levodopa if that was required according to their pretrial regimen.

Following baseline evaluation and overnight pre-trial levodopa/dopamine agonist washout on Day 1, participants assigned Sequence A received PF-06412562 (25 mg at ~09:00 h and 20 mg 4 h later) on Days 2 and 3 during Week 1, along with placebo (cellulose-filled capsule). During Week 2, these participants received encapsulated Sinemet (25/100 mg) 3–4 times (depending on pretrial regimen) 4 h apart on Days 2–3, along with placebo (Supplemental Table S2). The participants assigned Sequence B received carbidopa/levodopa in Week 1 and PF-06412562 in Week 2. On Day 4, all participants resumed pre-trial treatment and were discharged after demonstrating no significant complications. Blood levels of PF-06412562 were determined blindly (detailed in Supplemental Text S2).

Primary endpoints

The primary outcome was safety and tolerability of PF-06412562. Quantitative assessments included: vital signs, MDS-Unified PD Rating Scale dyskinesia scale-IV (MDS-UPDRS-IV), Columbia Suicide Severity Rating Scale (CSSRS), and routine safety tests (CBC, electrolytes, liver and kidney functions, ECG) at baseline and Day 4. Qualitative data included reports from participants, caregivers, nursing staff, and investigative team members.

Since hypotension was an adverse event observed with the first generation D₁ agonists such as dihydrexidine [25], and also with usage of approved D₂ agonists, supine blood pressure (BP) and heart rate (HR) were measured on Day 1 after admission (~16:00 h), on Days 2 and 3 at baseline prior to medication (~06:30 h), and at two and six hours after drug administration. In addition to supine vitals, cardiovascular autonomic testing was performed at these time points to assess cardiac sympathetic and parasympathetic tone. Although the original protocol included measurement of orthostatic vitals before and after drug administration, some of our advPD patients were unable to tolerate sitting or standing for prolonged periods of time at baseline (pre-test drug), and thus orthostatic measurements were not obtained.

For cardiovascular autonomic testing, participants were asked to rest in the supine position for at least 10 min, followed by 15 min of continuous BP and HR recordings. BP was measured continuously with finger photoplethysmography (Finometer FMS) and intermittently with an arm cuff via oscillometric methods (Philips SureSigns VS3). HR was measured using continuous three-lead ECG (Cardiocap/5, GE Healthcare). At the end of the 15-min recording period, subjects underwent respiratory sinus arrhythmia (HR response to deep breathing; 6 breaths/minute; normal response, maximum to minimum HR ratio >1.2) and cold pressor (hand in ice water for 1 min; normal response, increase in SBP >20 mm Hg) tests to assess the integrity of cardiac parasympathetic and sympathetic nerves, respectively. Resting HR variability, a measure of parasympathetic tone, was assessed in all subjects from continuous ECG recordings using time-domain spectral analysis methods [standard deviation of the average normal-to-normal intervals (SDNN) [26, 27].

Secondary endpoints

Two validated clinical global impression scales were modified a priori to capture both clinician and caregiver quantitative assessments of severity or change. For clinician assessment, the Clinician Global Impression of Severity (MD-CGI-S) first was used to assess baseline disease severity of the participant on Day 1. Second, the Clinician Global Impression of Change (MD-CGI-C) assessed clinical change from baseline (Day 1) using a 7-point Likert scale at two-hours post-dose. Lastly, the Overall Clinician Global Impression of Change Summary (MD-CGI-CS, Supplemental Table S3) is an overall clinician assessment of the entire week at the end of Day 3 using the same 7-point Likert scale. For caregivers, the scales were adapted from the clinician assessment tool as detailed in Supplemental Table S4. First, the Caregiver Global Impression of Severity (CG-CGI-S) was administered on Day 1 based on caregiver knowledge of the participant’s disease at home. Second, the Caregiver Global Impression of Change (CG-CGI-C) was assessed on the evening of Days 2–3 based on caregiver’s observation. Since Day 2 may be more confounded by excitement/noise/environment changes, we pre-determined that the evaluation performed at the end of Day 3 would be used for final analyses.

Statistical analysis

Primary endpoint analyses included all randomized subjects. Adverse events were summarized by MedDRA (ver. 21.2) criteria and severity (mild/moderate/severe). Vital signs and cardiovascular autonomic testing results were summarized using descriptive statistics. Laboratory results were evaluated immediately by physicians for clinical significance, and summarized at baseline and day-of-discharge. Secondary endpoint analyses included subjects who completed both study weeks. Data were summarized by treatment arms using descriptive statistics.

Results

Participants baseline clinical characteristics

A total of eight participants consented for the study and underwent a screening visit. Two were excluded due to unstable medical condition (ECG changes), or the need for medication adjustment after the screening visit. Six participants (detailed in Table 1) were on stable medication for greater than 60 days, and then were randomized to the study. All participants were in Hoehn and Yahr Stage >IV and not able to walk independently, two (001 and 008) were wheelchair/bed-bound and needed total assistance in transfer, four others (003, 004, 006, 007) required wheelchairs for transportation, and walkers, along with human assistance, for short distance walking or transfer.

Table 1.

Demographic, clinical history, and baseline data for the randomized participants

ID	Sex	Age of diagnosis	Disease duration	key medical, surgery, progression history	Current medications
001	M	37 y	19 y	Pallidotomy at 46 y Wheelchair use at 62 y PEG at 63 y	DA drugs: Parcopa Non-DA drugs: rivastigmine transdermal
003	M	60 y	18 y	PD at 60 y STN-DBS at 70 y Walker/wheelchair at 77 y	DA drugs: Sinemet R & CR Non-DA drugs: vitamin B12
004	F	56 y	15 y	STN-DBS at 62 y Walker use at 68 y Wheelchair use at 70 y	DA drugs: Sinemet R & CR, Rytary, Zelapar, Mirapex Non DA Drugs: dexlansoprazole, melatonin, midodrine, donepezil, memantine, clozapine, rimantadine, methylphenidate, venlafaxine fludrocortisdone
006	F	58 y	19 y	Walker use at 76 y Wheelchair at 77	DA drugs: Rytary, Sinemet Non-DA drugs: gabapentin, donepezil, lorazepam, quetiapine, melatonin, tramadol
007	M	53 y	23 y	Cane use at 73 y Wheelchair use at 75 y	DA drugs: Rytary, rasagiline Non-DA drugs: donepezil
008	M	42 y	31 y	On L-dopa until 70 y Bed-bound at 70 y	DA drugs: none* Non-DA drugs: none

^*:L-dopa discontinued due to severe drowsiness.

Abbreviations:F: Female; M: Male; Y: years; STN: Subthalamic nucleus; DBS: Deep brain stimulation; PEG: Percutaneous endoscopic gastrostomy; DA: Dopaminergic; Sinemet: carbidopa/levodopa; R: regular release; CR: controlled release.

One participant (001) had a history of pallidotomy that helped reduce levodopa-induced dyskinesia, and percutaneous endoscopic gastrostomy to assist nutrition status. Two participants (003 and 004) had DBS to the subthalamic nucleus, but remained Hoehn and Yahr Stage >IV even with stimulation on. The participant (008) with the longest disease duration of 31 y, benefited from levodopa and other anti-PD drugs earlier in his course of disease. He, however, experienced severe drowsiness,u and was unarousable without clear cause while on anti-PD medication including levodopa. His family elected to stop all medications including levodopa three years prior to this trial. He had erratic sleep-awake cycles prior to the study and no meaningful communication, but was able to swallow food when his mouth was stimulated.

Three participants (001, 006, 008) had severe baseline dehydration (SBP <100 and BUN/creatinine >20), with one (006) showing signs of pre-renal kidney failure [anuria, creatinine=1.34, estimated glomerular filtration rate (GFR)=38]. They were all treated with IV hydration on Day 1 (baseline) of each week per our pre-determined study procedures (see Table 2). Baseline vital signs and cardiovascular autonomic testing were obtained on Day 1 following any hydration procedures. There were no apparent overall differences in baseline vital signs or cardiovascular autonomic data between testing periods (Figure 3).

Table 2:

Administration of IV fluid (0.9% saline) and/or rescue levodopa usage for each participant

Study ID	Levodopa (SoC)			PF-06412562
	Day 1	Days 2–3	Day 4	Day 1	Days 2–3	Day 4
	Baseline	Test	Discharge	Baseline	Test	Discharge
001	0.9 % saline 500 mL	---	---	0.9 % saline 500 mL	---	---
003	---	---	---	---	---	---
004	---	Half tab of Sinemet R 25/100	One tab of Sinemet R 25/100	---	---	---
006	0.9 % saline 500 mL	0.9 % saline 500 mL	0.9% saline 500 mL	Trial terminated by safety board
007	---	---	---	---	---	---
008	0.9 % saline 1000 mL	---	---	0.9 % saline 1,000 mL	---	---

Figure 3. Vital signs and cardiovascular autonomic test results.

There were no apparent overall differences in baseline vital signs or cardiovascular autonomic data between testing periods.

All participants completed cold pressor testing, and four participants (003, 004, 006, 007) were able to follow breathing commands for sinus arrhythmia testing. Of these participants, three had abnormal cardiac sympathetic responsiveness (001, 006, 007: SBP changes: 9, −3 and −5 mm Hg, respectively), and one had abnormal cardiac parasympathetic responsiveness (006, sinus arrhythmia ratio=1.12), at baseline. One participant (006) with baseline impairment of cardiac sympathetic and parasympathetic nerve activity also exhibited reduced resting HR variability (SDNN=15 msec) when compared to the other participants (SDNN=37–106 msec).

Primary endpoints

No participants withdrew from the study voluntarily. No clinically meaningful laboratory test abnormalities were observed in either treatment periods. During the D₁ period, blood PF-06412562 concentrations were 424 ng/mL (SD=73; range 305–498 ng/mL), and were not detectable during the levodopa period (<0.55 ng/mL). One participant (004) required Sinemet 25/100 rescue on both Days 2 (1/2 tab) and 3 (one tab) for feeling more “off” during the SoC/levodopa period, but not the PF-06412562 period (Table 2). Most adverse events were rated as mild or moderate (Table 3).

Table 3.

Adverse events reported according to Grades, System Organ Class, and Preferred Terms

System Organ Class and Preferred Terms	Grade 1		Grade 2		Grade 3		Grade 4 or 5
Drug assignments**	LD	PF	LD	PF	LD	PF	LD	PF
Subjects completed	N=5	N=6	N=5	N=6	N=5	N=6	N=5	N=6
A. All events including Day 1 and one month after discharge
Any adverse event *	0	1	1	1	2	4	0	0
Any serious adverse event	0	0	0	0	0	1	0	0
Psychiatric disorders	0	1	1	0	0	1	0	0
Anxiety	0	1	0	0	0	0	0	0
Restlessness	0	0	0	1	0	0	0	0
Insomnia	0	0	0	0	0	1	0	0
Confusion	0	0	1	0	0	0	0	0
Hallucinations	0	0	0	1	0	0	0	0
Vascular/Automatic disorders	0	1	0	0	2	3	0	0
Hypotension	0	1	0	0	0	1	0	0
Hypertension	0	0	0	0	2	2	0	0
Gastrointestinal disorders	0	0	1	0	0	0	0	0
Gastrointestinal pain	0	0	1	0	0	0	0	0
Musculoskeletal disorders	1	0	0	1	0	0	0	0
Pain	1	0	0	1	0	0	0	0
Nervous System disorders	1	0	0	2	0	0	0	0
Headache	0	1	0	1	0	0	0	0
Fall	1	0	0	0	0	0	0	0
Dizziness	0	0	0	1	0	0	0	0
B. Events after test drug each week (including Days 2 and 3, and after discharge)
Any adverse event *	0	0	0	2	2	3	0	0
Any serious adverse event	0	0	0	0	0	1	0	0
Psychiatric disorders	0	1	1	1	0	0	0	0
Anxiety	0	1	0	0	0	0	0	0
Restlessness	0	0	0	1	0	0	0	0
Insomnia	0	0	0	0	0	0	0	0
Confusion	0	0	1	0	0	0	0	0
Hallucinations	0	0	0	1	0	0	0	0
Vascular/autonomic disorders	0	0	0	0	2	3	0	0
Hypotension	0	0	0	0	0	1	0	0
Hypertension	0	0	0	0	2	2	0	0
Gastrointestinal disorders	0	0	0	0	0	0	0	0
Gastrointestinal pain	0	0	0	0	0	0	0	0
Musculoskeletal disorders	1	0	0	1	0	0	0	0
Pain	1	0	0	1	0	0	0	0
Nervous System disorders (total)	1	0	0	2	0	0	0	0
Headache	0	0	0	1	0	0	0	0
Fall	1	0	0	0	0	0	0	0
Dizziness	0	0	0	1	0	0	0	0

^*Number of patients with an adverse event, summarized for each treatment group.

^**LD = levodopa/carbidopa; PF = PF-06412562; PT = preferred term; SOC = system organ class. MedDRA version used was 22.01

For each PT, adverse event per patient are counted only once according to the most severe grade. AEs were assessed for severity using the Common Terminology Criteria for Adverse Events (CTCAE) with Grade 1–5 classification (1:mild; 2:moderate; 3:severe; 4: life-threatening; 5:death)

There were no significant differences in BP or HR, with a similar mild to moderate hypotension and reflex tachycardia observed following PF-06412562 or levodopa administration (Figure 3). Three participants (003, 004, 008) with normal baseline cardiac sympathetic responsiveness did not change following PF-06412562 or levodopa. Similarly, three participants (003, 004, 007) with normal cardiac parasympathetic responsiveness at baseline showed no differences in response to PF-06412562 versus levodopa. There were also no significant differences in HR variability between drugs.

One participant (006) who exhibited severe sympathetic impairment and mild parasympathetic impairment at baseline experienced symptomatic hypotension and was treated with IV hydration on Day 3 of the PF-06412562 week. The participant was discharged on Day 4 with improved laboratory results (creatinine=0.98, BUN=24, and estimated GFR=55), normal ECG findings, and appeared clinically stable. The participant, however, required emergency room visits at 1, 13, 15, and 24 days and subsequent hospitalizations to treat hypotension after Week 1 discharge. Although the participant and caregiver wanted to proceed to study period 2, we followed the recommendation of the DSB to discontinue their participation.

Secondary endpoints

As seen in Table 4, no subjects were rated worse (negative scores) in clinical global impression of change than baseline during PF-06412562 periods by either clinicians or caregivers, whereas one clinician and three caregivers rated subjects worse during the SoC/levodopa period. Overall caregiver scores favored PF-06412562 in four participants (001, 003, 004, 007) and levodopa in one participant (008) (Table 3). Clinician impression of change scores were more diverse, with two (003, 004) favoring PF-06412562, two (001 and 008) levodopa, and one (007) equivocal.

Table 4.

The overall clinical impression of change at the end of Day 3

Participants’ Study ID	Levodopa (SoC) period			PF-06412562 period
	Week*	Clinician**	Caregiver***	Week	Clinician	Caregivers
001	1	2	8	2	1	22
003	2	0	−1	1	3	27
004	1	−1	−4	2	3	21
007	1	0	−4	2	0	13
008	2	2	32	1	0	2

^*Week when the drug was administered/tested

^**Clinician rating on a 7 point Likert scale (−3=marked worsening; 0=no change; 3=marked improvement) (see details in Supplemental Table S3)

^***Caregiver rating is the sum [−51 (marked worsening) to[51 (marked improvement)] of 17 items listed in Supplemental Table S4 with each item on a 7 point Likert scale as above.

One participant (008) was ‘awakened’ during Week 2 when given Sinemet. After extensive discussions with the research ethicist, we disclosed this finding to the family for their consideration. The family decided to resume Sinemet, but the ‘awakened’ state did not recur.

Discussion

Recent therapeutic development for PD has focused on disease modification [28–30], but unfortunately remaisn unsuccessful. Current symptomatic treatments have allowed many PD patients to have life expectancies 15 years or more after diagnosis [31], although quality-of-life decreases with continuing disease progression. Due to limited pharmacological options, management of advPD patients transitions to one of support and palliative measures [32]. As the first drug trial targeting PD patients in advanced-to-end-stages of disease, this study elucidated many of the challenges, as well as potential solutions. Equally important, this study demonstrated the feasibility of studying this population, and provided preliminary data that PF-06412562, a partial D_1/5 agonist, was tolerated in a small cohort of advPD patients.

Although past and ongoing clinical trials have avoided inclusion of patients with advPD, this study showed that patients and their family are willing to participate in drug trials. Between July 16, 2018 (IRB approval) to the end of study enrollment on December 30, 2019, we enrolled eight subjects at a single site, and received additional request for participation after the enrollment deadline. This enthusiasm and willingness for participation underscores the need for better drug treatment options for this population.

No participants withdrew from the study voluntarily. and PF-06412562 was well tolerated in advPD patients. This is consistent with previous clinical trials of milder-stage PD [23]. Based on results of prior studies with other agents, a main focus was assessment of potential for hypotension. Indeed, the only selective D₁ agonist approved clinically, fenoldopam, is indicated for emergency acute hypertensive crises. Fenoldopam, however, does not pass the blood brain barrier, and therefore is not centrally active [33, 34]. Although having antiparkinson effects, the first generation full D₁ agonist dihydrexidine has been associated with hypotension requiring intensive care unit admission for some subjects [25]. We suspected that this side effect may have been due, in part, to rapid intravenous administration for the study [25], as slower subcutaneous infusion of dihydrexidine was well tolerated, achieved measurable blood levels [35], permitting use in studies of cognition efficacy in psychiatric patients [36, 37].

All current dopamine replacement therapy lowers BP somewhat, and in some patients precipitates orthostatic hypotension [38]. It was thought that dopamine receptor stimulation induces orthostatic hypotension by venous and arteriolar dilation. Consistent with this notion, we found similar BP lowering (Figure 3) in both the PF-06412562 and levodopa arms even in the supine position, an effect greater on the second treatment day (Day 3). Unfortunately, we were not able to determine if PF-06412562 affects orthostatic BP, as we originally intended, since some participants (e.g. subjects 001 and 008) were not able to sit/stand for prolonged periods of time. Future studies therefore may need to incorporate head-up tilt table testing to evaluate the degree and tolerability of BP drops. Nonetheless, PF-06412562 was well-tolerated with symptomatic hypotension only observed in one participant (006) with pre-existing dehydration and acute pre-renal kidney injury.

In the current study, we found that half of participants (3/6) had normal cardiac sympathetic integrity, and 3/4 of participants who completed sinus arrhythmia testing had intact cardiac parasympathetic integrity. There were no differences in cardiac sympathetic and parasympathetic tone or responsiveness, suggesting PF-06412562 or levodopa did not adversely impact cardiovascular autonomic function. Of note, the participant (006) who developed symptomatic hypotension had the most significant sympathetic and parasympathetic dysfunction. This suggests that screening for cardiovascular autonomic function will be valuable in future trials. In addition, to protect against any potential cardiovascular vasodilatory effects, we speculate that pre-treatment of advPD patients with therapies to increase sympathetic tone (such as the norepinephrine prodrug droxidopa) may be useful in future studies with PF-06412562 or other D₁-like agonists.

Since dehydration is common in advPD patients, we did not exclude advPD patients with dehydration in the current study. Instead, we put into place pre-determined study criteria to protect this subset of patients against potential serious cardiovascular outcomes such as symptomatic hypotension. We recognize this may alter resting cardiovascular tone and drug responsiveness. We did not observe any apparent differences, however, in vital signs or cardiac sympathetic and parasympathetic tone between participants who did or did not receive hydration support. Of note, the participant (006) who developed symptomatic hypotension also had pre-renal kidney dysfunction. Future studies may need to consider including baseline renal dysfunction, even secondary to dehydration, as an exclusion criterion (Table 2).

A secondary goal of our study was to provide pilot data for possible efficacy of D₁ agonists in advPD. Because of the complex, multi-domain dysfunction of advPD patients, we chose to include a caregiver evaluation to improve the possibility of detecting meaningful changes. A single clinician rater was not feasible throughout the study, so the data are from two clinician raters and one caregiver for each of the participants. The overall results from the caregiver assessments favored PF-06412562 over levodopa, but the clinicians’ impressions found no difference. Although one would predict that expert clinicians would provide more objective evaluations, the clinicians’ scores may suffer from inter-rater variability in judging the severity of such advPD patients. Conversely, the caregivers were experienced with the subject they evaluated, and might be more sensitive to changes. Future studies should emphasize the use of a single clinician rater for a specific participant as much as possible, and consider integrating caregiver evaluations as an integral part of data collection.

In summary, although physicians, patients, and caregivers have recognized the many unmet needs of advPD patients for decades, these patients essentially have been excluded from therapeutic clinical trials. This study provides proof-of-concept that therapeutic drug trials are feasible in advPD, and that D_1/5 agonists are tolerated. Because its potential to provide benefit in advPD, future studies are warranted to include a larger sample size, longer treatment duration, additional safety measurements (such orthostatic BP with tilt table), and patient-centric meaningful outcome assessments that include caregivers’ perspectives.