Peptide drug products

Peptide Microstructure Characterization

Peptide implants, microspheres, oral tablets and lyophilized products release according to a microstructure that a certificate of analysis never shows. We image that structure in 3D, measure it, and predict release.

Start Your Project See what we can measure →

Peptide development teams

The Peptide Teams We Work With

RoleChallengeWhat we deliver

Peptide formulation scientistsMonths of in vitro release testing for every sustained release formulation, and lyophilized cakes that collapse or reconstitute slowly.Release predicted from the measured structure with ivisLab and dissoLab, and the 3D pore and cake structure measured in I2S.

Manufacturing and process engineers (CMC)Batch to batch variability when spray drying, freeze drying or polymer mixing scales up.A digital representation that links your process parameters to the 3D microstructure attributes that set quality.

Regulatory and generic bioequivalence leadsShowing Q3 sameness to the reference product for generic peptide depots.Non-destructive 3D Q3 microstructure data for the comparisons FDA asks complex generics to make.

Peptide application areas

Where Peptide Structure Decides Performance

Three areas across the peptide lifecycle where a product succeeds or fails on its internal structure.

Application areaThe challengeWhat we do

Long-acting injectables, microspheres and implantsPLGA microspheres, in situ forming depots and solid implants

3D render of a FIB-SEM scan of a microsphere showing its internal microstructure — FIB-SEM volume of a microsphere interior, rendered in 3D

A sustained release peptide depot can require months of release testing for each formulation change, and a late failure restarts the study.

Non-destructive micro-CT and FIB-SEM imaging map drug distribution, polymer shell thickness and internal pore connectivity. ivisLab then predicts release from the measured structure in days, not months.

Lyophilized cake quality and reconstitutionFreeze dried peptide cakes reconstituted at the point of care

3D render of a lyophilized cake with the pore space left by ice crystals in blue — Lyophilized cake rendered in 3D. Blue marks the pore space left by ice crystals.

Lyophilized peptide cakes are susceptible to pore collapse, shrinkage and vial to vial variation in reconstitution time.

Voxel level measurement of cake porosity, pore size distribution and wall thickness, allowing freeze drying cycles to be compared by the cake structure each produces.

Solid oral peptides and permeation enhancersOral peptide tablets formulated with a permeation enhancer

Micro-CT cross section of a tablet, the left half as scanned in grayscale and the right half segmented by AI into colored phases — Micro-CT slice of a tablet, raw on the left and segmented by AI on the right

Oral peptide tablet performance depends on the microscale distribution of the peptide relative to the permeation enhancer.

3D mapping of the proximity between peptide domains and permeation enhancer, and of whether that arrangement is consistent from batch to batch.

Images are representative digiM scans of each dosage form and do not show peptide products.

Implant drug loading

Same Peptide and Polymer Different Release

We imaged fifteen long-acting implants made by hot melt extrusion, all the same peptide in the same polymer. At 15% drug load the implant held zero-order release for four months, as designed. At 25%, about 78% of the dose was released within the first seven days, and at 20%, about 56%.

Percolation theory predicts burst release at 15% as well. At 15% by weight the drug takes up 17.2% of the volume, above the point where a dispersed phase links into a connected network in three dimensions. The measured drug network was 90% interconnected, consistent with that prediction.

Yet the 15% implant showed no burst. Its network was connected but poorly conductive, with a permeability 56 times lower than that of the 25% implant. Resolving this required 3D reconstruction of the implant interior and flow modeling through the drug network.

The practical percolation threshold

We define the practical percolation threshold as the drug loading at which the network begins to govern release. It does not appear on a certificate of analysis. It can be measured at day zero on implants already manufactured, months before a release study would reveal it.

Read the study →Assess your loading series →

Cumulative release by drug load

Measured in vitroPredicted from images

Release in the first seven days

15% drug load	About 4%, then zero-order release over four months
20% drug load	About 56%
25% drug load	About 78%

Same peptide, same polymer, hot melt extruded implants. Points are in vitro release tests. Lines are release predicted from FIB-SEM images of the as-made implants, after diffusivity and polymer degradation were calibrated against measured release. Values read from Zhang et al., Journal of Pharmaceutical Sciences, 2021, Figure 10f.

Drug volume

17.2%

of the implant volume is drug at a 15% load by weight, above the textbook percolation threshold

Network connectivity

90%

of the drug network at 15% load is interconnected, consistent with percolation theory

Permeability

56×

lower permeability than the 25% implant, which is why the 15% implant held its release

Characterization capabilities

What We Can Measure and Predict

Each study uses only the methods its question needs.

Image and segmentFrom one unit or your own images

Micro-CTNon-destructive 3D imaging at half a micron
FIB-SEMResolution to 30 nanometers for finer structure
AI segmentationDrug, polymer, pores and excipients each labeled through the volume
3D visualizationThe full volume or individual phases

MeasureEvery phase, through the whole volume

Volume percentage of each phase
Size distribution of each phase, from pores to API particles
Equivalent diameter
Spatial distribution
Porosity and pore connectivity
Tortuosity
Permeability of the drug network
Practical percolation threshold
Polymer shell thickness
Shell to core porosity gradients
Distance between phases, such as peptide to enhancer
True density

Predict and compareFrom the measured structure

Release predictionivisLab and dissoLab, to prioritize formulations before release testing
Q3 sameness evidenceYour measured structure set against miQ3, our database of reference listed drug microstructure metrics
Synthetic microstructuressGAN, our generative AI that augments limited image data with realistic synthetic microstructures

Peer-reviewed research

Publications

Burst prediction

R² 0.82

between predicted and measured burst release rate over the first seven days, from day zero images of 15 implants

Prediction time

3 to 6 days

to predict a six month release test: 1 to 2 days imaging, 1 to 2 days AI image analysis, 1 to 2 days release prediction

Release profiles

profiles for implants designed to release over four to six months, produced in a few weeks

Zhang et al. and Nagapudi et al., Journal of Pharmaceutical Sciences, 2021

Journal of Pharmaceutical Sciences, 2021 · two papers

Long-acting peptide implants imaged at half a micron and again at 30 nanometers

Fifteen hot melt extruded implants. The papers measure the drug network, find the practical percolation threshold, and predict burst release from day zero images.

Read the publication →Journal of Controlled Release, 2023

Thousands of PLGA microspheres per batch measured non-destructively in 3D

Eight compositionally equivalent batches. Microspheres 20 µm and larger were structurally uniform within each batch, and structural variation between batches correlated inversely with release rate.

Read the publication →

Limitations

Imaging at half a micron cannot resolve features below 1 to 2 µm, so it is not the tool for nanoscale structure.
The authors note that imaging alone cannot fully characterize a microsphere product.
The microsphere structure to release correlation has three exceptions among eight batches.
The closest implant release prediction required calibrating two transport parameters against measured release data from one implant.

Microstructure CRO study

Study Workflow

Each study begins with one unit and one defined question. Every stage lists its deliverables and turnaround.

Scope call

We define the dosage form and the study question with you. Together they determine the imaging method, resolution and cost.

One call

Baseline Scan

One unit imaged, segmented and quantified, with a written assessment of whether the microstructure accounts for the observed behavior.

10 to 14 working days

Performance CRO

Release prediction, comparison against a reference, or a lyophilization cycle comparison across several lots or formulations.

3 to 6 weeks

Software Subscription

Run I2S, ivisLab and dissoLab in your own lab, with training from the scientists who apply them to active programs.

Ongoing

A six month in vitro release test can be predicted in three to six days: one to two days of imaging, one to two days of AI image analysis, and one to two days of release prediction.

Zhang et al., Journal of Pharmaceutical Sciences, 2021

Pharmaceutical imaging software

The Software Behind Every Study

We apply these tools to active peptide programs. The same software is available for license to run in house.

ScreenFirst look at the API

segHT PLM

Detects crystalline and amorphous domains in polarized light microscopy images of the API.

→

MeasureImage and quantify structure

digiM I2S

Turns an image stack into segmented, quantified structure.

→

GenerateAugment limited data

sGAN

Generates realistic microstructures when imaging data is limited.

→

PredictFrom structure to release

ivisLab

Predicts release for implants, microspheres and depots from the measured structure.

→

dissoLab

Predicts dissolution for oral products from particle structure.

→

CompareTest against reference

miQ3

Reference listed drug microstructure metrics, measured with one consistent method, as the benchmark for generic comparison.

→

The Difference Between Two Batches Is Inside the Dosage Form

A study can start from one unit or an existing image stack. We return the grayscale and segmented images, the measured structure of every phase and 3D visualizations for internal review.

Request a Live Demo