Preparation of Frozen Non-Human Primate Fetal Islets for Combined Single Nuclei RNA-Sequencing and ATAC-Sequencing, and Bulk Metabolomics

Summary

This protocol describes procedures to isolate high-quality nuclei from frozen non-human primate pancreatic islets for use in single-nucleus simultaneous RNA sequencing and ATAC sequencing while preserving the bulk cytosolic fraction for metabolomic analyses from the same samples.

Abstract

One challenge in studies using tissue collected from multiple cohorts is avoiding batch effects when preparing for large-scale multi-omic experiments, such as combined single-cell RNA sequencing and metabolomics. The method in the current study utilizes flash-frozen pancreatic islets from fetal non-human primates collected over a span of two years for input into single-nucleus RNA sequencing and ATAC sequencing assays. The cytosolic fraction generated during nuclear extraction was retained for downstream capillary electrophoresis-mass spectrometry and subsequent metabolite quantification. This method allows for bulk analysis of metabolites that contribute to the changing transcriptomic and epigenomic landscapes within experimental conditions. It is applicable to many tissue types and maximizes the amount of information that can be extracted from samples that are not readily available. As the contribution of metabolism to the establishment of cellular identity via epigenetic modifications becomes more appreciated, techniques that allow for identifying the contribution of metabolites in specific cell types are timely and necessary.

Introduction

The pancreatic islets of Langerhans are critical for the management of glucose tolerance throughout the lifespan. Failure to properly tune blood glucose levels via the glucose-lowering hormone, insulin, or the glucose-raising hormone, glucagon results in Diabetes Mellitus¹. Modifiable risk factors to minimize diabetes risk include exercise and diet. In addition, it is becoming clear that maternal diet during pregnancy also has an effect on offspring susceptibility to diabetes in adulthood^2,3. Thus, methods are needed to analyze the effects of maternal diet and other exposures, such as medications used during pregnancy, on the insulin-producing beta cells in model organisms during development. For example, islets from fetal non-human primate (NHP) offspring can be evaluated for the effects of developmental exposures. Similar to humans, NHPs typically have singleton births and vary in their physiological response to Western-style diet feeding⁴. While the gestation and fetal development of NHPs shares many similarities with humans, making them a highly relevant species for metabolic programming studies, challenges to performing comparative bioinformatic assays on NHP offspring include a restricted mating season and intermittent sample collection. Therefore, approaches that allow for long-term storage prior to sample processing and analysis are ideal. This protocol describes the preparation of samples for single nucleus RNA-sequencing, ATAC-sequencing, and bulk cytoplasmic metabolomics.

Studies by others in the islet biology field have demonstrated overlap in the genetic signatures observed in samples prepared via single-cell RNA Sequencing and single-nucleus RNA sequencing^5,6. Both methods allow for the identification of transcripts that contribute to distinguishing features amongst various cell types within the islet population. Through the use of gentle detergents on flash-frozen tissue, intact nuclei can simultaneously be used for ATAC sequencing, which allows for the determination of accessible chromatin regions. Combining single-nucleus RNA-sequencing with ATAC-sequencing allows for better characterization of gene regulatory networks⁷.

The preparation of samples for metabolomics has traditionally been limited by the need for a high sample volume input for liquid chromatography. In addition, careful consideration must be given to the buffers used during sample preparation to avoid false NMR peaks in mass spectrometric data. Given the low number of samples that are challenging and costly to acquire in primate studies, allocation of islets from different offspring toward transcriptomic versus metabolomic experiments can be difficult. Thus, this protocol makes use of the normally discarded cytosolic fraction generated during cellular fractionation for single-nucleus experiments. This method allows for the simultaneous characterization of transcripts and open chromatin regions that define cellular subsets in fetal islets while quantitatively measuring the metabolites produced within islets (Figure 1). This protocol is adapted from a demonstrated protocol published by 10x genomics for the isolation of nuclei from flash-frozen embryonic mouse brain tissue (version CG000366 Rev D⁸)⁷.

Protocol

Islet retrieval was conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of the Oregon National Primate Research Center (ONPRC) and Oregon Health and Science University and was approved by the ONPRC IACUC. The ONPRC abides by the Animal Welfare Act and Regulations enforced by the United States Department of Agriculture (USDA) and the Public Health Service Policy on Humane Care and Use of Laboratory. The protocol herein was applied to islets from fetal Rhesus macaques that were necropsied on gestational day 145 (Rhesus macaque gestation is typically 173-175 days). The details of the reagents and the equipment used for this study are listed in the Table of Materials.

1. Islet isolation

1. Excise fetal pancreata from surrounding tissues⁹ and place it in ice-cold phosphate-buffered saline.
2. Remove the pancreas from PBS and inflate it with collagenase P (0.6 mg/mL) via cannulation of the pancreatic duct using a 33 G rodent brain cannula injector.
3. Divide the pancreas into six sections with sterile surgical scissors and digest for 27-30 min at 37 °C in collagenase solution.
4. Prepare RT media by combining 10% FBS and 100 U/mL penicillin/streptomycin into RPMI 1640.
5. Pour off collagenase and add 10 mL of RT media.
6. Gently shake for 1 min to disrupt the remaining connective tissue.
7. Remove undigested material by straining through a 500 µm filter.
8. Wash the islets twice with 50 mL of RT media.
9. Incubate the islets in bovine DNase I (0.8 U/mL) in a 37 °C water bath for 10 min.
10. Culture the islets overnight in RT media with bovine DNase I (0.4 U/mL) at 37 °C with 5% CO₂.
11. Ship islets in specialized packaging at ambient temperature with temperature control gel packs.

2. Sample banking for long-term storage

1. Assess islets for size using a scaled reticle and note the degree of acinar and ductal tissue in preparation.
2. Handpick islets with an automated pipette into a 60 mm dish containing 5 mL of 2.8 mM glucose DMEM.
   NOTE: It is imperative to minimize the amount of acinar tissue and ductal tissue that will be stored in the final islet preparation. Details of ideal islets for use are reported in a previous protocol⁹.
3. Pick islets into a 1.5 mL tube and spin at ~300 x g for 5 min at 4 °C.
4. Resuspended the islets in calcium and magnesium-free 1x PBS.
5. Spin for 5 min at ~300 x g at 4 °C. Discard the supernatant.
6. Repeat steps 2.3-2.5 twice more.
7. Flash freeze the islets in a 1.5 mL tube for 10 s in liquid nitrogen.
8. Store the pellet at -80 °C until the day of nuclear isolation.

3. Cellular fractionation for isolation of nuclei and cytoplasmic fraction

1. Preparation of buffers for nuclei extraction (Table 1).
   1. Ensure that the bench, pipettes, and reagents are RNase-free.
   2. Prepare the wash, concentrated lysis, lysis dilution, and diluted nuclei buffers according to Table 1.
   3. Prepare 0.1x lysis buffer in a 5 mL tube using the lysis dilution buffer to dilute the concentrated lysis buffer.
2. Tissue homogenization
   1. Remove the samples from -80 °C freezer and immediately place them on ice.
   2. Add 500 µL of 0.1x lysis buffer to the tissue pellets and incubate for 3 min on ice.
   3. Using a disposable RNase-free pellet pestle, gently homogenize the tissue 15 times on ice.
   4. Incubate the samples on ice for 10 min.
   5. Add 500 µL of wash buffer to the homogenate and gently pipette on ice.
   6. Prime a 30 µm pre-separation filter by adding 300 mL of wash buffer to the filter over a 5 mL tube.
   7. Filter the 1 mL cellular homogenate using the 30 µM pre-separation filter into the 5 mL tube.
   8. Spin the filtered homogenate at 500 x g for 5 min in a swinging bucket centrifuge at 4 °C.
3. Collection and storage of supernatant for metabolomics
   1. Remove 1 mL of the supernatant from step 3.2.8 with a pipette and add to a 1.5 mL cryovial on ice.
   2. Immediately freeze the supernatant at -80 °C. This supernatant will be used for metabolomics.
4. Preparation of nuclei for single nucleus RNA- and ATAC-sequencing
   1. Gently resuspend the nuclei pellet by adding 1 mL of wash buffer to the pellet dropwise. Slowly pipette 5 times.
   2. Spin the washed pellet at 1000 x g for 5 min in a swinging bucket centrifuge at 4 °C.
   3. Repeat steps 3.4.1 and 3.4.2 once more.
   4. After the final wash, resuspend the nuclei in 1 mL of wash buffer. Keep nuclear preparation on ice.
   5. Aspirate 10 µL of nuclei suspension with a pipette and mix with 10 µL of 0.4% Trypan blue by gently pipetting.
   6. Add 10 µL of the nuclei and Trypan blue mixture to an automated cell counter slide with a pipette and place it in the instrument.
   7. Determine nuclei concentration in nuclei/µL.
      NOTE: Unlike cellular preparations for single-cell RNA sequencing, automated cell counters will detect nuclei as dead cells. No more than 95% of this preparation should be live on automated cell counters. If more than 5% of the nuclear population is live, the nuclei suspension can be filtered an additional time through the 30 µm pre-separation filter and spun for 5 min at 1000 x g at 4 °C. Nuclei concentration can then be recalculated as outlined in step 3.4.6.
   8. Based on the desired targeted nuclei recovery, resuspend nuclei to the necessary concentration using 1x nuclei buffer and proceed to single nucleus RNA and ATAC sequencing library preparation¹⁰.

4. Preparation of cytosolic fraction for capillary electrophoresis- mass spectrometry (CE-MS)

1. Combine 80 µL of cytosolic fraction collected in step 3.3.1 with 20 µL of nanopure water containing internalized standards.
2. Proceed with CE-MS and analysis as previously described^11,12,13.

Results

The quality of the sequencing outputs is highly dependent on the quality of the nuclear input. Nuclear blebbing and damage to the outer nuclear membrane will result in ambient RNA and DNA contamination in the analysis, which will introduce noise in the data. A pure nuclear extract will have minimal evidence of nuclei encapsulated with cytoplasmic components (Figure 2A) or overdigestion, evidenced by bursting nuclei (Figure 2B). An ideal nuclear input will contain intact nuclei (Figure 2C).

Details of the analysis of single nucleus RNA- and ATAC-sequencing data are extensively provided elsewhere^5,14. Representative clusters of cellular populations from a fetal NHP nuclear preparation are illustrated via UMAP in Figure 3A. Metabolomics data is readily analyzed using open-source software, as previously described¹⁵. Generated spectra can be visualized, and enriched metabolites in islets of a fetus from a WSD-fed dam can be compared to those from a CD-fed dam (Figure 3B-F).

Figure 1: Overview of protocol for isolation of nuclei and cytosolic fraction. Fetal pancreatic islets were banked and stored at -80 °C until the day of the assay. Nuclei were isolated and utilized for single nucleus RNA- and ATAC-Sequencing. The cytosolic fraction following the first wash of nuclei isolation was retained for capillary electrophoresis-mass spectrometry. Please click here to view a larger version of this figure.

Figure 2: Demonstration of proper nuclei input into single nucleus multiomics assay. Examples of (A) Nucleus encapsulated by cytoplasm, (B) Bursting nucleus with compromised nuclear membrane, and (C) Intact nucleus with rounded morphology. Scale bars: 10 µm. Please click here to view a larger version of this figure.

Figure 3: Representative data from combined single nucleus multiomics and bulk metabolomics. (A) UMAP clustering of cellular subtypes within isolated fetal NHP islets based on gene expression. Quantitative estimation of the glycolytic metabolites, (B) Dihydroxyacetone phosphate and (C) Glycerol-3-phosphate in the cytosolic fraction from isolated fetal NHP islets when the dam was fed a WSD (mWSD) or a CD (mCD). Maternal diet may also differentially regulate metabolism via the (D) citric acid cycle (2-oxoglutarate), (E) amino acid metabolism (creatine), and by altering (F) cellular redox status (glutathione) in fetal islets. Please click here to view a larger version of this figure.

Table 1: Buffer formulations for isolation of nuclei. Detailed components of nuclei, lysis, lysis dilution, and wash buffers are shown. If necessary, dilute reagents to the stock concentration before preparing the buffer. Please click here to download this Table.

Discussion

Historically, the methods of nuclear isolation for single nucleus RNA- and ATAC-sequencing have utilized buffers containing detergents that are incompatible with liquid chromatography-mass spectrometry (LC-MS), a common technique used for metabolite quantification^16,17. The current protocol allows for the combination of transcriptomic and metabolomic techniques from the same sample preparation by using a gentle detergent for cellular fractionation, and following this fractionation with capillary electrophoresis, a technique that requires a smaller sample input relative to liquid chromatography for the separation of metabolites¹⁸.

The most critical aspect of this protocol is the preservation of the nuclear membrane during nuclear isolation to avoid the degradation of genetic material and reduce contamination by ambient nucleotides and macromolecules. Sample exposure times to lysis buffer can be modulated to avoid over-digestion of tissues and nuclei damage. Additionally, quick freezing of the islets during sample banking and the supernatant containing the cytosolic fraction is imperative to avoid the degradation of genetic material and metabolites. All samples must be collected from the same media formulation (2.8 mM glucose DMEM in this study) to avoid the confounding effect of altered metabolism due to nutrient availability in culture media.

A limitation of this combined metabolomic approach is that some metabolites may not be captured due to the detergent used for nuclei isolation. Therefore, for a more complete metabolite analysis, this approach could be complemented with analyses from frozen tissues and isolating metabolites using non-organic wash buffers and LC-MS. A recommended supplemental approach is targeted spatial metabolomics via MALDI-MS to localize metabolic changes within organs¹⁹. Additionally, the present technologies for evaluating transcript variability and chromatin accessibility within the same nucleus are limited due to shallow sequencing depth for ATAC sequencing. In data analyses, investigators will only be able to evaluate open chromatin regions within broad cellular clusters rather than specific subclusters within the dataset.

As the use of tissues that are not readily available, particularly those from humans and non-human primates, becomes more frequent, it is imperative to maximize the amount of information that can be extracted from a biological replicate. The multi-omic approach described above allows investigators to correlate genetic signatures with peak availability around key promotors of cellular identity while also considering how cellular metabolism may have contributed to such changes. Identifying the metabolites that are more abundant within cellular preparations will give clues to the specific epigenetic modifications that would drive the specification of cellular subtypes to a particular lineage. Indeed, new protocols have been established for integrating the findings of transcriptomic and metabolomic experiments, albeit from different tissue sources²⁰. The significance of this protocol is that chromatin accessibility, transcriptional changes, and metabolite enrichment can be evaluated from a single sample input. Future applications of this technique will allow for better characterization of disease states in tissues acquired from sparse sample populations, such as tissues from patients with rare diseases. In the islet biology field, the characterization of gene regulatory networks and the metabolic signatures that correlate can provide critical insight as to mechanisms of preserving beta-cell identity and function to avoid the onset of diabetes.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
33 G rat brain cannula injector	Protech International	8IC315IS5SPC	For duct cannulation and pancreas inflation
Penicillin/Streptomycin	Fisher	15-140-122	For RT media
FBS	Millipore/Sigma	F4135-500ML	For RT media
RPMI 1640	Thermo Fisher Scientific	11-875-135	For RT media
Collagenase P	Sigma Aldrich	11213865001	For digestion of pancreas
Bovine Dnase I	Sigma Aldrich	11284932001	For RT media and isolation
Dextrose	Fisher	D16-1	Final concentration will be 2.8 mM
Calcium Chloride	Fisher	C4901-500G	Final concentration will be 0.5 mM
HEPES	Gibco	15630-080	Final concentration will be 10 mM
Bovine Serum Albumin (BSA) Fraction 5	Research Products International	A30075100	Final concentration will be 0.5 mg/mL
Dulbecco's Modified Eagle Medium	Gibco	11966025	For handpicking islets
20x Nuclei Buffer	10X Genomics	2000153/2000207
Digitonin (5%)	Thermo Fisher Scientific	BN2006	Digitonin may need to be warmed to 65 °C to dissolve white precipitate
MACS SmartStrainers (30 µM)	Miltenyi Biotec	130-098-458
MACS BSA Stock Solution (10%)	Miltenyi Biotec	130-091-376
IGEPAL CA-630	Sigma Aldrich	i8896	Prepare a 10% stock using MilliQ water
Trizma Hydrochloride Solution, pH 7.5	Sigma Aldrich	T2319-1L
Sodium Chloride Solution, 5 M	Sigma Aldrich	S6546-1L
Magnesium Chloride Solution, 1 M	Sigma Aldrich	M1028-100mL
Sigma Protector Rnase inhibitor	Sigma Aldrich	3335402001
DTT	Research Products International	D11000-5.0
Rnase-Free Disposable Pellet Pestles	Fisher Scientific	12-141-368
Tween 20	Fisher Bioreagents	BP337-500	Prepare a 20% stock using MilliQ water
Nuclease-Free Water	Promega	PR-P1193
DNA LoBind Tubes (5 mL)	Eppendorf	30108310
Cryovials (1.5 mL)	VWR	20170-225
Posi-Click Tubes (0.6 mL)	Denville	C-2177
Trypan Blue (0.4%)	Thermo Fisher Scientific	T-10282
Dual-Chamber Cell Counting Slides	Bio-Rad	1450011
MiiliQ Ultrapure Water System	Millipore/Sigma	7003/05/10/15
5 kDa cut-off filter	HMT Metabolome Technologies
Internal Standards	HMT Metabolome Technologies
P10, P20, P200, and P1000 Pipettes	Eppendorf	2231000602
Pipette Tips, 10 µL	VWR	76322-528
Pipette Tips, 1000 µL	Thermo Fisher Scientific	21-236-2A
Pipette Tips, 20 µL	Genesee Scientific	23-404
Pipette Tips, 200 µL	USA Scientific	1120-8810
Phase Contrast Microscope	Olympus	BX41-PH-B
Countess II Automated Cell Counter	Thermo Fisher Scientific	AMQAX1000
DPBS (1x)	Gibco	14190-144
Allegra X-30R Centrifuge	Beckman-Coulter	B06315
RNase-ZAP	Sigma Aldrich	R2020