Optimization and use of near infrared imaging to guide lymph node collection in rhesus macaques (Macaca mulatta)

Abstract Background Identification of lymph nodes (LNs) draining a specific site or in obese macaques can be challenging. Methods Indocyanine Green (ICG) was administered intradermal (ID), intramuscular, in the oral mucosa, or subserosal in the colon followed by Near Infrared (NIR) imaging. Results After optimization to maximize LN identification, intradermal ICG was successful in identifying 50–100% of the axillary/inguinal LN at a site. Using NIR, collection of peripheral and mesenteric LNs in obese macaques was 100% successful after traditional methods failed. Additionally, guided collection of LNs draining the site of intraepithelial or intramuscular immunization demonstrated significantly increased numbers of T follicular helper (Tfh) cells in germinal centers of draining compared to nondraining LNs. Conclusion These imaging techniques optimize our ability to evaluate immune changes within LNs over time, even in obese macaques. This approach allows for targeted serial biopsies that permit confidence that draining LNs are being harvested throughout the study.


| INTRODUC TI ON
One of the key advantages of macaque models is the ability to perform in-depth tissue sampling in ways that are not possible in human patients. To this end, we had previously developed minimally invasive (MIS) laparoscopic techniques to permit sampling of key immunologic sites, [1][2][3][4] which permit further characterization of tissue responses to infectious diseases and immunology in the macaque model. One important site that we have sampled in the past is mesenteric lymph nodes (MLNs) which are important reservoir sites for SIV and are immunologically distinct from peripheral lymph nodes (PLN) 3,5,6,7,8,9,10 due to their exposure to microbes and microbial products coming from the GI tract. These sites had previously been difficult to sample serially, but our MIS techniques have been used to sample them at up to eight timepoints during the course of a study, with the main obstacle to successful collection being the difficulty to identify MLN in obese animals.
Sampling of peripheral lymph nodes in macaque models can also present challenges, with the need to sample sites repeatedly combined with factors such as obesity; LNs can be hard to find resulting in missed time points. As macaque studies are typically only powered sufficiently to detect differences when all samples are evaluated, missed time points can impact the ability to achieve meaningful and statistically significant results. Additionally, for events that are localized in nature, such as immunization at a specific site, it can be important to ensure that the LNs collected are actually draining the site of interest. Drainage patterns can vary by individual 11,12 and nondraining LN (LNs not draining the site of interest) from the same PLN site (axillary, inguinal, submandibular, etc.) can easily be inadvertently collected unless targeted collection is performed.
To address issues associated with accuracy of lymph node collections, we have optimized a dye-guided lymph node (DGLN) collection technique for use in the macaque model. This technique utilizes the dye indocyanine green (ICG) which is visualized with near infrared (NIR) imaging to allow for 1. Targeted collection of LNs from specific sites, and 2. Identification of LNs in obese animals where they were not identifiably using standard methods. Here, we present the DGLN technique for identification of MLN in obese animals and the identification of LNs draining the site of immunization in both the oral mucosa (intraepithelial or IEp inoculation) and bicep muscle (IM inoculation). Draining LNs were compared to nondraining LNs from these sites (axillary LNs for IM and submandibular LNs for oral mucosa IEp) following immunization and significant differences in replication and levels of key immune cells within germinal centers were identified.

| Humane care guidelines
All animals were housed, cared for, and used in accordance with the policies of the Institutional Animal Care and Use Committee (IACUC) at the ONPRC, an AAALAC-accredited institution, which abides by the USDA Animal Welfare Regulations, the Public Health Service Policy on Humane Care and Use of Laboratory Animals and the Guide for the Care and Use of Laboratory Animals. All enrolled macaques were SPF (serologically negative for SIV, simian type D retrovirus, and Macacine herpes 1) prior to the start of the study and were captive-born at ONPRC. All animals were singly or socially housed in accordance with their IACUC protocol. Throughout the study, all animals were uniformly fed Purina LabDiet 5000 (Purina Mills International), daily nutritional enrichment items (grains, fruits, or vegetables) and had ad libitum access to water. In addition to social housing and nutritional enrichment, animals received environmental enrichment in the form of toys, music/radio, and television daily. All animals were observed for health or behavioral concerns at minimum twice daily by research, veterinary, and/or husbandry staff. Animals requiring euthanasia for study or clinical endpoint purposes were sedated with an intramuscular injection of 20 mg/kg ketamine HCl (Ketathesia™, Henry Schein Animal Health) followed by an overdose of intravenous pentobarbitol solution, in accordance with the AVMA Guidelines for the Euthanasia of Animals.
Nineteen rhesus macaques (15 males/4 females), ranging from 3 to 16 years of age and 5.6-12.1 kg were used to optimize ICG administration for fluobeam PLN detection. Dose volume, number of injection sites, and duration post-administration were analyzed to optimize LN signal identification for inguinal and axillary LN sites.
Collection of MLN and PLN using ICG guidance in obese rhesus macaques was performed as part of their primary study once standard methods had failed to identify an MLN and/or PLN at that timepoint.
Thirteen obese rhesus macaques (7 males/6 females) ranging from 7 to 17 years old and 7.5-16.1 kg were evaluated; 7/13 were used for PLN, and 9/13 were used for MLN detection. A total of 2 rhesus (females, 10-11 years old, and 7-8.5 kg) macaques (n = 1/group) were injected with ICG labeled 10 014_F8.V3.N7.SOSIP (natively HIV envelope folded trimer) either by IM injection in the biceps brachii or IEp injection in the oral mucosa. A boost injection at the same site was performed 4 weeks after the initial immunization, and the draining LNs (submandibular for the oral site, axillary for the IM) were collected 2 weeks after the boost. Unbound ICG was also injected at the site of immunization prior to surgery to ensure that sufficient signal was present in the draining LNs for identification by NIR.
Nondraining (no fluorescent signal) LNs from the same site (axillary for biceps brachii IM and submandibular for oral mucosal vaccine administration) were collected at the same time for comparison to the draining LNs.

| Administration of SOSIP trimer
10 014_F8.V3.N7.SOSIP Envelope protein was expressed in HEK293 cells and purified by GNA lecting chromatography as previously described. 13 The sequence is derived from an env sequence isolated from an HIV-1+ subject (10014) that developed broadly neutralizing antibodies approximately 3 years post-infection (cite 25 122 781).
For conjugation, protein was mixed with NH-Reactive ICG and Reaction Buffer and incubated for 10 min at 37°C. The reaction was stopped with WS buffer and filtered, and the labeled protein was diluted into phosphate-buffered saline (PBS). Labeling efficiency was determined by assessing absorbance at 280 nm and 800 nm and calculating the absorption coefficient. ICG-labeled Env vaccine formulation was prepared with the following adjuvants so that each injection contained 100ug Env trimer, 100ug R848 (Invivogen), 10ug MPLA (Invivogen), and 1% Alhydrogel (Invivogen). The labeled SOSIP trimer was injected in a volume of 0.1 ml/site into the oral mucosa of both the left and right cheek pouch. For IM, a volume of 0.25 ml was injected into the right and left biceps brachii.

| Biopsy procedure
Biopsies were performed as described previously. Once prepared for surgery, sites were draped and all procedures were performed using aseptic technique. Incisions were made with a #15 surgical blade and biopsy sites were closed in 2-3 layers depending on location, using 4-0 absorbable monofilament suture. Axillary and inguinal sites were closed in two layers (subcutaneous and skin), while submandibular and laparoscopic port sites were closed in three layers (muscle, subcutaneous, and skin). All skin closure was performed using an intradermal pattern, followed by application of cyanoacrylate tissue adhesive (Vetbond™ Tissue Adhesive, 3 M). Instillation of local anesthesia along the incision was also performed.

| Cell counts
PLN and MLN were obtained as described above and were processed then stained for flow cytometric analysis. Lymph nodes were collected in R10 media (RPMI-1640 with 10% sterile filtered Newborn Calf Serum, 1% penicillin-streptomycin, 1% L-glutamine, 1% sodium pyruvate, and 50 nM/mL of 2-Mercaptoethanol) at biopsy and were then processed using a pestle and screen for CD-1 (Sigma-Aldrich, S1020-5EA) in a tissue culture dish. Processed tissues were strained through a 40 μm filter into a 50 ml conical, centrifuged at 700 rcf for 10 minutes, and resuspended in R10 before obtaining the cell yield using a Horiba ABX Pentra 60 C+ hematology analyzer.

| Statistical analysis
All data are shown as standard deviation (SD) of the mean.
Optimization results were analyzed using paired t-tests for comparison of signal intensity at different times within an animal and unpaired t-tests for comparison of volume and number of sites using GraphPad Prism (San Diego, CA, USA) version 9.0 for Windows.
Results were considered statistically significant when the p value was less than .05.

| Optimization of ICG for axillary/inguinal lymph node identification
ICG was injected at a concentration of 0.625 mg/ml at various locations in the axillary and inguinal areas of macaques as either a single site or at two sites. This concentration was sufficient to allow for identification of multiple LNs down the lymphatic draining chain of nodes and had a mild quenching effect at the center of the injection bleb when imaged (see Figure 1). Together, these findings demonstrated that there was sufficient signal present for identification of LNs and that further increases could actually lessen the signal intensity due to quenching. 14 For identification of nodes, ICG injection at two sites was superior to one site with a signal intensity of threefold greater at 5 min (p = .032), 3.3-fold greater at 10 min (p = .007), and 2.6-fold greater at 15 min (p = .013) when one site on the limb (arm or leg) and one site on the chest/abdomen were used for axillary and inguinal, respectively, at a 0.2 ml volume of ICG (see Figure 2). We determined that a 0.2 ml volume/site presented the optimal volume to achieve signal both in terms of time and signal intensity (p = .047 at 5 min and p = .018 at 15 min). We then assessed the signal intensity in the nodes at different times post 2 site ID administration of 0.2 ml of ICG and determined that 5 min was optimal through visual assessment.
We then assessed the numbers of LNs at each site with and without fluorescence at necropsy for animals that underwent the 0.2 ml/site x 2 site ICG injections. Animals had between 2 and 14 nodes present per axillary/inguinal site with an average of 77% (50%-100%) of the nodes present at the site demonstrating detectable fluorescent signal using the Fluobeam800, and all animals had at least 2 nodes that were positive at each site. Thus, optimization of our dye-guided lymph node (DGLN) collection technique resulted in 0.2 ml/site x 2 site ICG injections with imaging performed at least 5 minutes post ID administration for PLN detection.

| Use of ICG for PLN and MLN collection in obese macaques
ICG as optimized for PLN DGLN collection was used in a total of 7 obese rhesus macaques after traditional methods for identification of axillary/inguinal LNs had failed. PLN were identified and collected in 100% (7/7) using DGLN resulting in an average lymphocyte yield of 31.1 × 10 6 /animal (0.8-120 × 10 6 ). For MLN, the same concentration of ICG was used for DGLN but was injected at ~0.4 ml/site at 2-3 sites subserosally in the colon (see Figure 3). After ~5-10 min, MLN were visualized using the Stryker1688 and were collected in 100% (9/9) obese rhesus macaques where traditional methods of MLN identification had failed resulting in an average lymphocyte yield of 54.1 × 10 6 /animal (8-185 × 10 6 ).

| Comparison of draining to nondraining lymph nodes
NIR imaging was utilized to target LNs draining the site of an HIV Env SOSIP immunization in two rhesus macaques, one immunized via IM injections and a second via an IEP injections to the oral mucosa. Two doses of vaccine were administered 8 weeks apart.
Two weeks after the second vaccination, either axillary (in IM vaccinated) or submandibular (in orally vaccinated) lymph nodes were obtained. NIR DGLN imaging was utilized to identify both nondraining ( Figure 4A) and draining ( Figure 4B) lymph nodes in each of these areas (designated ICG-and ICG+, respectively, Figure 4C).

Increased size and number of germinal centers (GCs) was observed
in the draining LNs ( Figure 4A compared to B). In addition, the

| DISCUSS ION
We demonstrated the utility of our optimized DGLN technique using ICG, an FDA-approved fluorescent dye that has both an extremely good safety profile and minimal potential to impact immune responses, for identification of both PLN and MLN in rhesus macaques.
This resulted in a 100% rate of identification of LNs that could not be identified by traditional methods. Avoiding missed timepoints is key not only for ensuring that missed samples from unique animals do not skew results, but also to ensure that sufficient numbers are present to achieve statistical significance. Most macaque studies are only powered sufficiently to allow for statistical comparisons when the data set is complete, and missing samples that require elimination These DGLN techniques can be employed either using a labeled target (antibody, antigen, virus, etc.) [15][16][17] or by injecting unbound ICG at the same site 18,19 or both using the signal from the labeled target to guide administration of the unbound ICG which in turn serves to increase the signal intensity in the draining LN. Identification of actual draining LNs is key for any localized response, 20-23 such as an immunization, local infection, or early local viral spread after a mucosal exposure. Early events in macaque models of SIV have been difficult to assess due to inter-animal variability and/or the need for serial sacrifice at early timepoints that do not permit following an animal to determine disease progression and kinetics. 10,24,25 Serial comprehensive sampling, especially using targeted DGLN collections, could permit evaluation of differences in responses that lead to different viral kinetics or differences in progression to systemic infection vs local control and eventual elimination overtime. Following responses, overtime from preinfection to setpoint viremia will be essential for optimizing responses to vaccines and therapeutics aimed at early intervention to reduce reservoirs and setpoint viremia in SIV models designed to inform clinical trials in HIV patients. In addition to providing better data, avoiding unnecessary animal intense serial sacrifice studies will also result in a reduction in the number of animals required to address an experimental question. Given that the MIS techniques have very low complication rates (typically <1%) 4 and do not require the use of anti-inflammatory or antibiotic medications the main limitations to their use are equipment and surgeon training.
The use of NIR imaging is ideal for DGLN identification given the long wave length which leads to greater tissue penetration when compared to other fluorescence imaging techniques. ICG is ideally suited as a DGLN contrast agent for use in NIR detection of draining LN given the long (>800 nm) wavelength of emission and the corresponding depth (~1 cm) of tissue fluorescence signal penetration. [26][27][28] We observe minimal to no autofluorescence in this range, allowing for excellent signal-to-noise ratios. Because of the lipoprotein binding capacity of ICG and the high protein content of lymph, ICG accumulates in the lymphatic pathways and LNs. 29 Additionally, given its small size (~775 daltons) ICG is extremely unlikely to generate immune responses. These characteristics also make ICG superior to other techniques involving nonfluorescent inks/dyes that can be mistaken for natural pigments, such as hemosiderin, blood accumulation in the LN, or the presence of tattoo ink from animal identification, and which have been shown to have inflammatory properties. [30][31][32] The movement of ICG from the site of injection to the primary draining LN occurs very quickly, with 5 minutes representing the ideal balance between wait time and signal intensity based on the optimization work we have done.
Additionally, if the ICG is administered early in the process of preparing the site for surgery, there is no additional time required for the procedure as it takes ~5 minutes to shave, prep, and drape the site in most cases.
In conclusion, we have developed NIR DGLN techniques for identification of PLN and MLN, providing a powerful tool that can be used to study immunity in macaques with a high degree of accuracy and specificity. Applications that analyze lymph node-specific responses are heavily dependent on accurate recovery of immunologically active lymphoid tissue and thus will greatly benefit from the incorporation of NIR-guided LN identification. More broadly, because ICG can be coupled to many types of molecules, this technique is compatible with virtually any approach aimed at identifying trafficking of biologics or compounds to the lymph nodes, creating the opportunity to selectively and accurately analyze outcomes of LN-specific interventions.

ACK N OWLED G M ENTS
This work was funded by grant R01DE026336 to DNS and DLS. This work was supported by the Office of the Director, National Institutes of Health (3 U42 OD023038-03S1), and Oregon National Primate Research Center NIH Core Grant (P51OD011092). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

CO N FLI C T O F I NTE R E S T
The authors do not have any conflicts of interest in regards to this publication.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.

E TH I C S S TATEM ENT
This study was performed in accordance with the ethical policies of the journal, and in compliance with Guide for the Care and Use of