Alector : PK/PD modeling of progranulin elevation in blood and CSF to support AL101 dose selection and phase 2 study design

PK/PD Modeling of Progranulin Elevation in Blood and CSF to

™

Support AL101 Phase 2 Study

Massimiliano Germani1, Amitkumar Joshi2, Carey Hines2, Robert Lai1, David Roth3, Michael Ward4, Lovingly Park4, Balasubrahmanyam Budda4

1GSK Neuroscience, Stevenage, UK 2PPD, part of Thermo Fisher Scientific, Wilmington, NC, USA 3GSK Research and Development, Collegeville, PA, USA 4Alector, Inc., South San Francisco, CA, USA

Background

• The PK/PD profiles were simulated to build the dose-response curve in terms of average change from baseline of PGRN in CSF

Figure 3. Estimated Effect of BW on AUCtau in Response to Weight- Adjusted vs Flat Dosing Regimens

Figure 5. PGRN Elevation in CSF From Baseline at Trough (A) and as an Overall Average (B)

• Progranulin (PGRN) is a glycoprotein that is encoded by the GRN

gene, regulated by sortilin-mediated endocytosis, and known to play

a vital role in many cellular processes, such as inflammation, wound

repair, lysosomal function, and neurodegeneration1,2

• GRN mutations resulting in downregulation of PGRN have been

linked to the development of neurodegenerative disorders, such as

Alzheimer's disease (AD)3

• Rare GRN loss of function mutations are reported in clinically

diagnosed AD patients; the common variant rs5848 is associated

with a ~15% decrease in plasma PGRN and has been identified as a

Figure 1.  PK/PD Model Structure and Parameterization

Drug input (SC)

Depot	Drug
input (IV)

K12a

Body weight adjusted was compared with flat dosing (same amount for 67kg subject)

AUCtau

Between subject variability defined by the predicted percentiles, 5th and 95th meaning 90% of the patients are expected to be within the range defined by the dotted lines

Local regression curves prediction intervals

PGRN elevation change from baseline at trough (%) in CSF

DOSE

elevation change from (%) in CSF

q4w: typical value (median) and quartiles (25th and 75th percentiles)

q8w: typical value (median) and

genetic determinant of AD2

• Given the neurotrophic actions of PGRN, increasing PGRN levels

may provide an appropriate therapeutic approach for individuals with

neurodegenerative conditions such as AD2,4

Peripheral

Central

(PK)

(PK)

K21c

K13b		Linear
CSF
		(PK)

Linear

Body weight adjusted dosing

Flat dosing

PGRN average baseline

DOSE

quartiles (25th and 75th percentiles)

q12w: typical value (median) and quartiles (25th and 75th percentiles)

• AL101 is a human immunoglobulin (Ig) G1 monoclonal antibody

designed to bind to sortilin and inhibit interaction between PGRN

and the sortilin receptor, thereby elevating PGRN levels in blood and

cerebrospinal fluid (CSF) and slowing the rate of decline in individuals

Decrease elimination of PGRN

Plasma (PGRN)

Decrease elimination of

PGRN

CSF

(PGRN)

BW

500 simulations per dose

AUCtau, area under the curve at over the dosing interval; BW, body weight.

CSF, cerebrospinal fluid; PGRN, progranulin; q4w, every 4 weeks; q8w, every 8 weeks; q12w, every 12 weeks.

Conclusions

with AD5

• The safety, tolerability, pharmacokinetics (PK), pharmacodynamics

(PD), and bioavailability of AL101, when administered as single or

multiple intravenous (IV) or subcutaneous (SC) doses, were evaluated

in healthy individuals in a phase 1 study6

Objective

• To simulate a dose-response (in terms of average PGRN elevation during the dosing period in plasma and CSF) profile based on the phase 1 study results

Methods

• The PK/PD model (Figure 1) was developed based on IV data from phase 1 study5,6 and used to build a dose-response profile that tested the effects of AL101 while changing dose level and frequency
• Simulation studies establish a mathematical relationship between AL101 plasma and CSF concentrations and the expected change in PGRN plasma and CSF levels
• The final model included body weight (BW) as a covariate on clearance and volume of distribution in serum AL101, and baseline plasma PGRN as a covariate on the maximum effect term on plasma PGRN degradation. Once validated the model was used for simulating different clinical scenarios
• Simulations were performed considering the patient populations:
• • BW and baseline PGRN levels were extracted from a normal distribution from the phase 1 data, and the model assumed no difference in PK/PD between healthy volunteers (phase 1 study) and AD patients (simulated phase 2 study)
• 1000 simulations were conducted to estimate the effects of either a high dose administered every 4 weeks (q4w) IV (Figure 2A) or low dose administered q4w IV (Figure 2B) on AL101 concentrations in plasma and CSF and the corresponding percent change from baseline in PGRN levels in plasma and CSF (data not shown)

aK12 = transition rate from central compartment to peripheral compartment bK13 = transition rate from central compartment to CSF compartment cK21 = transition rate from peripheral compartment to central compartment

CSF, cerebrospinal fluid; IV, intravenous; PD, pharmacodynamics; PGRN, progranulin; PK, pharmacokinetics; SC, subcutaneous.

Figure 2. PK/PD Simulations: Estimated Concentrations of AL101 in Plasma and CSF of AD Patients in Response to High (A) or Low (B) Doses Based on q4w IV Dosing

Simulated AL101 concentrations

A		High Dose	B		Low Dose
	10000			10000
	1000			1000
ug/mL	100		ug/mL	100	CSF
			Typical value (median)
			5th and 95th percentiles
	10			10	Quartiles
Conc		Conc	(25th and 75th percentiles)
1		1	Quartiles
			Plasma
					Typical value (median)
	0.1			0.1	(25th and 75th percentiles)
					5th and 95th percentiles
	0.01			0.01
	0.001	14 28 42 56 70 84 98 112 126 140 154 168 182 196 210	224	0.001	14 28 42 56 70 84 98 112 126 140 154 168 182 196 210 224
	0	0
		Time (days)			Time (days)

AD, Alzheimer's disease; CSF, cerebrospinal fluid; IV, intravenous; PD, pharmacodynamics; PK, pharmacokinetics; q4w, every 4 weeks.

Results

• Providing the estimated parameters, variability in model parameters, residual errors, and random extracted covariates (ie, BW and PGRN baseline values), the PK/PD model was used to simulate AL101 (Figure 2) concentrations in plasma (red) and CSF (blue)
• BW was detected as a statistically significant covariate in the PK/PD model. However, its impact on area under the curve at steady state (AUCss), over the dosing interval (AUCtau), and maximum concentration at steady state (Cmax) [data shown in log-scale] suggested mostly similar exposure-distribution between flat and weight-based dosing and minimal or no impact on the safety margin (Figure 3)

• Figure 4 shows the relationship between dose (AL101 q4w IV) and

response (steady-statebaseline-corrected CSF PGRN concentrations)

and Figure 5 demonstrates the impact of changing the dosing

frequency considering q4w, q8w, and q12w

• Two IV dose levels will be considered for the phase 2 study. A high

dose q4w IV is proposed as the top dosing regimen to reach and

maintain the maximum expected PGRN elevation in CSF

- A dose that exceeds the high dose proposed here would not be expected

to yield an improved PD effect

- Increasing the dosing interval from q4w would impact the PGRN

concentration profile and lead to lower CSF PGRN levels with major

impact at the end of the dosing period

Figure 4. AL101 Average Expected Dose Response Based on Varying Dose Levels Administered q4w IV (2000 simulations per dose level)

CSF - change from baseline (%)

Steady state

Low Dose

High Dose

DOSE q4w

CSF, cerebrospinal fluid; q4w, every 4 weeks.

• PK/PD characterization and simulations supported the transition to flat dosing compared to weight-based dosing and justified the administration of two dose levels at q4w intervals
• The high dose q4w is expected to be the minimal dose which would allow obtaining the maximum PGRN elevation in CSF and maintain it for the entire dosing period, whilst the lower dose level is expected to be the median effective dose (ED50) in terms of average PGRN elevation from baseline in CSF
• Additional simulation testing could allow for further investigation regarding the effects of dosing frequency and the impact of utilizing different routes of administration
• The PK/PD model will be integrated with emerging data to increase the confidence in parameter estimation and confirm and inform potential covariate effects

References

1. Chitramuthu BP, Bennett HPJ, Bateman A. Progranulin: a new avenue towards the understanding and treatment of neurodegenerative disease. Brain. 2017;140(12):3081-3104.
2. Rhinn H, Tatton N, McCaughey S, Kurnellas M, Rosenthal A. Progranulin as a therapeutic target in neurodegenerative diseases. Trends Pharmacol Sci. 2022;43(8):641-652.
3. Wang XM, Zeng P, Fang YY, Zhang T, Tian Q. Progranulin in neurodegenerative dementia. J Neurochem. 2021;158(2):119-137.
4. Hu F, Padukkavidana T, Vaegter CB, et al. Sortilin-mediated endocytosis determines levels of the frontotemporal dementia protein, progranulin. Neuron. 2010;68(4):654-667.
5. Ward M, Paul R, Maslyar D, et al. A first-in-human study of the anti-sortilin antibody AL101. Poster presented at: 14th Clinical Trials on Alzheimer's Disease Conference; November 9-12, 2021; Boston, MA.
6. Ward M, Yeh FL, Park L, et al. Repeat IV and SC dosing of the anti-sortilin antibody AL101. Poster presented at: 15th Clinical Trials on Alzheimer's Disease Conference; November 29-December 2, 2022; San Francisco, CA.

Disclosures

Massimiliano Germani, Robert Lai, David Roth are all employees and shareholders of GSK. Lovingly Park and Balasubrahmanyam Budda are employees and shareholders of Alector, Inc. Mike Ward is an independent consultant of Alector, Inc. and Athena Bioscience, LLC, and a shareholder of Alector, Inc. Carey Hines and Amitkumar Joshi are employees of PPD and may hold equity in PPD.

Acknowledgments

Medical writing services were provided by Scient Healthcare Communications and funded by Alector, Inc. We thank the site staff, participants, and their families for participation in the clinical trial.

Presented at the Annual Meeting of the American College of Clinical Pharmacology | September 10-12, 2023 | Bellevue, WA, USA