Evaluation of 415-nm Blue Light and Hypochlorous Acid on the Rhesus Macaque (Macaca mulatta) Cranial Implant Microbiota

Abstract

Cranially-implanted macaques are a crucial model for neuroscience research. Implant complications include abscesses, meningoencephalitis, and implant–tissue margin infections. Antimicrobial overuse has increased bacterial resistance, risking macaque health and complicating treatment of infections. This project aimed to assess the antimicrobial effects of blue light (∼415 nm) and hypochlorous acid (HOCl) treatment of cranial implant margins. Blue light exerts antibacterial effects via the induction of reactive oxygen species. We hypothesized that exposing cranial implant margins to a commercially available blue light device followed by HOCl treatment would improve clinical appearance and decrease bacterial burden as assessed by aerobic/anaerobic culture and tissue margin microbiota analysis (decreased α and β diversity and altered taxonomic composition). Eight rhesus macaques were exposed to 6 min of blue light followed by 0.024% HOCl solution three times weekly for 4 wk. Swabs for microbiota analysis and bacterial cultures were collected before and 24 h after the last treatment session. Control microbiota swabs were collected from a separate implant margin area only exposed to HOCl. All animals tolerated the blue light exposure but had varied improvement in margin clinical appearance. The most common bacteria identified on culture were Staphylococcus aureus (n = 8), β-hemolytic streptococcus (n = 8), and Corynebacterium ulcerans (n = 6). Microbiota analysis of the 16S rRNA V4 gene region demonstrated many anaerobic operational taxonomic units in addition to the aerobic species cultured, highlighting limitations of culture-based methods. All animals had unique microbiota taxonomic profiles with a mean of 84 operational taxonomic units and a median Shannon diversity index of ∼2.6. No significant differences were found between treatment groups, α diversity, or β diversity before and after the study. The effectiveness of blue light therapy likely relates to the device power and depth of penetration into the tissue margin. While safe, future work is needed to optimize the dose and delivery methods of light-based therapies.

Introduction

Macaques are commonly used in behavioral neuroscience research due to their forward-facing, binocular color vision and ability to perform cognitive tasks.^1,2 Neuroscientists study the brain by using cranial implants, including titanium headposts, cephalic recording chambers (CRCs), and microelectrode arrays to record neuronal activity during awake behavioral trials.³ Infectious complications of chronic cranial implants include implant/skin margin infections, meningoencephalitis, osteomyelitis, and brain abscess.^4–6 The treatment of cranial implant complications is hindered by antimicrobial penetration through the blood–brain barrier, the presence of biofilm coating implant surfaces, and the increasing prevalence of antimicrobial resistance in research NHPs.^7–12 Chronic bacterial colonization has been documented in association with the cranial implant margin,¹³ in association with titanium alloy headposts,⁴ and in addition within CRCs.^10–12 Previous work has identified that the microbiota of CRCs is most similar to the skin microbiota at the cranial implant margin, as compared with the oral microbiota and fecal microbiota.¹⁰ Current guidelines for care of cranial implants in NHPs are described in detail for both maintaining CRCs as well as the implant/wound margin.⁷ For maintenance of implant margins, the Association of Primate Veterinarian guidelines recommend gentle removal of hair and loose crusts and cleaning with various solutions including sterile saline, 0.05% chlorhexidine solution, 1% to 2% povidone-iodine solution, Dakin solution (0.5% sodium hypochlorite), 0.012% hypochlorous acid (HOCl), or 1.5% to 3% hydrogen peroxide.⁷ Despite the many methods described, novel adjunct therapies are still sought to decrease bacterial burden, minimize infectious complications, avoid additional selective pressure for antimicrobial-resistant bacteria, and maintain cranial implants for animals on neuroscience studies.

Blue light at a 415-nm wavelength is hypothesized to exert an antimicrobial effect via inactivation of bacterial catalase¹⁴ and/or the photoexcitation of endogenous chromophores (porphyrins and flavins), resulting in the generation of reactive oxygen species.^15–17 Efficacy of blue light has been reported against multiple bacterial species, including Staphylococcus aureus, Acinetobacter baumannii, Escherichia coli, Pseudomonas aeruginosa, Enterococcus faecium, Porphyromonas gingivalis, Fusobacterium nucleatum, and Prevotella intermedia in both planktonic and biofilm states.^18–20 Commercially available, FDA-cleared devices are marketed for use in humans as antiacne devices for direct application to the face and body. Multiple clinical trials in humans have demonstrated safety and clinical improvement with utilization of these devices.^21–24 Given the current utilization of disinfectant solutions with oxidizing properties (for example, dilute sodium hypochlorite, HOCl, and hydrogen peroxide), blue light therapy could provide synergistic results for cranial implant margin maintenance when used in combination with these products.

This study aimed to evaluate the safety and use of an FDA-cleared blue light device on the skin-implant microbiota in chaired macaques. To determine the antimicrobial effects (or lack thereof) of the blue light device we compared both traditional culture methods and 16S microbiota results before and after treatment. We hypothesized that a commercially available, FDA-cleared blue light device could be used safely in chaired macaques at the implant/skin margin. We also hypothesized that exposure of the implant/skin margin to a blue light device in combination with a veterinary HOCl solution would affect the microbiota. Specifically, we hypothesized that catalase-producing bacterial strains such as S. aureus and Corynebacterium spp. would be more susceptible to oxidative damage from HOCl following blue light exposure and we would see a decrease in α and β diversity as compared with a control site treated only with HOCl. Finally, we hoped to see clinical improvement in the appearance of the implant/skin tissue interface.

Materials and Methods

All procedures were approved by the Harvard Medical Area Standing Committee on Animals (IACUC). Eight rhesus macaques (Macaca mulatta) (1 female, 7 males) actively enrolled on separate IACUC-approved neuroscience protocols were used for this study. The sample size was determined by the number of animals who were actively on study for their neuroscience protocols and/or were expected to be temperamentally compatible sitting in a primate chair outside of performing standard computerized behavior tasks. Macaques ranged in age from 3 to 17 y (mean age 9.9 y). Macaques are screened annually and have remained serologically negative for B virus, simian T-lymphotropic virus, SIV, and simian retrovirus. The demographics of animals enrolled are presented in Table 1. Macaques are housed in an AAALAC-accredited facility meeting or exceeding standards of the Animal Welfare Act and Guide for the Care of Use of Laboratory Animals (8th ed.). Macaques were maintained on a 12-h light/12-h dark cycle, fed a commercially formulated Old World primate chow (Purina Lab Diet 5045 and/or 5LG1; Richmond, IN) supplemented with fresh produce, dried fruits, and various rotating dietary enrichment. Some animals received water and fluids on a scheduled basis for behavioral task performance per their IACUC-approved primary neuroscience protocols. All macaques were chaired and head-fixed according to their routine neuroscience protocols during sampling and blue light treatment and provided with dietary rewards, human interaction, and/or video entertainment based on the preference of the individual animal. We excluded a priori any animals that had received antimicrobial treatment within 30 d of study initiation. All macaques enrolled in the study had a minimum of 56 d since their previous surgical procedure and antimicrobial treatment. CRCs are not used at our institution; all animals have headposts alone, or in combination with implanted microelectrode arrays. After microarrays are implanted, the craniotomy is sealed with polymethylmethacrylate acrylic (PMMA) and the external microarray connector is protected with a cap. Because there is no exposure of the craniotomy site, sterilization of cylinder caps is not required as it would be for a traditional CRC. The frequency of cleaning implant margins depends on the individual animal, and appearance of the implant margin and was not directly controlled for. Implant margin cleaning regimens vary by animal and implant appearance but typically include a combination of warm water, gentle debridement, HOCl spray, and occasional 1% to 2% povidone-iodine solution. During the study period, povidone-iodine solution was not used for implant margin maintenance, and mechanical debridement was minimized to avoid confounding effects. Due to the small number of animals enrolled, no randomization or blinding was performed.

Table 1.

Demographics of Macaca mulatta enrolled in the blue light study

ID	Sex	Age (y)	Cranial implant(s)	PMMAa	Previous surgery	Days since previous antimicrobial treatment	Animal source
S	M	5.5	Headpost + arrays	+	Array implant	567	Born in-house
L	M	6	Headpost + arrays	+	Array implant	574	Born in-house
B	M	12	Headpost + arrays	+	Array implant	84	Born in-house
P	M	17	Headpost + arrays	+	Array implant	98	Born at primate center
T	F	3	Headpost + arrays	+	Array implant + craniotomy	185	Born in-house
G	M	16	Headpost	+	Headpost placement	543	Born at primate center
A	M	10	Headpost	−	Array explant	56	Transfer from another institution
D	M	10	Headpost; erosion site	−	Array explant	62	Transfer from another institution

All animals with arrays have implanted chronic microelectrode arrays; no CRCs are used at our institution.

^a This column refers to the presence or absence of polymethylmethacrylate as a component of the cranial implant/skin margin.

Sample collection

Bacterial culture swabs (BD BBL CultureSwab Plus 220116; Becton, Dickinson and Co., Franklin Lakes, NJ) were collected around the implant margin and submitted for aerobic and anaerobic culture to a commercial veterinary diagnostic laboratory (IDEXX BioAnalytics, North Grafton, MA). Bacterial identification was performed using MALDI-TOF, and antimicrobial susceptibility testing was performed via minimum inhibitory concentration (MIC) testing per laboratory protocols. For microbiome analysis, flocked nylon swabs (Puritan HydraFlock; Puritan Medical Products, Guilford, ME) were collected at 2 separate implant margin sites, designated as a blue-light + HOCl site and a HOCl-only site. Macaques A and D cranial implants consisted only of titanium headposts without external PMMA. For macaque D, an area of skin erosion with a visible titanium screw caudolateral to the headpost was used for the blue light + HOCl treatment, and the craniolateral aspect of the headpost margin (on the opposite side of the erosion) was used for the HOCl only site. For macaque A, opposite sides of the headpost margin were used for the blue light + HOCl treatment and HOCl-only treatment. For all other macaques, different areas of the skin/PMMA margin were used for the 2 sampling sites. Swabs were placed in preautoclaved microcentrifuge tubes and frozen at −20 C° until submission. Samples were identified by the animal’s ID with prestudy samples numbered 1 and poststudy samples numbered 2.

Treatment

Hair at the exposure site was trimmed gently using a mini-shaver. A commercially available acne treatment device (reVive Light Therapy; LED Technologies, Largo, FL) was used to expose a designated implant margin area to 415-nm blue light for 2 consecutive 3-min treatment periods, for a total of a 6-min exposure time (Figure 1A, B). The device was positioned to be in direct contact with the exposure site, or as close to direct contact as possible, given the geometry of the cranial implant present at the exposure site. Following the blue light treatment, the exposure site and the control site were sprayed with 0.024% HOCl (Vetericyn). The surface of the light was sanitized using accelerated hydrogen peroxide disinfectant (Peroxigard; Virox Technologies, Oakville, ON, Canada) in between animals, and outer nitrile gloves were changed in between animals. Animals were treated 3 d per week (Mondays, Tuesdays, and Fridays) for a total of 12 sessions. The order of the macaques treated within a treatment day was determined by the schedule of the investigating laboratories and not controlled. Hair at the exposure site was trimmed as needed during the study. The day following the last treatment session, culture swabs of the blue light + HOCl site were collected along with swabs for microbiome analysis of both the blue light + HOCl site and HOCl-only site.

Figure 1.

(A) Image of the blue light device used for the study displaying the surface shape and array layout. (B) Representative image of application of the blue light device to the cranial implant margin.

Sequencing and microbiota analysis

Swabs collected for microbiome analysis were shipped on dry ice overnight to the Alkek Center for Metagenomics and Microbiome Research at the Baylor College of Medicine. DNA was extracted using the Qiagen DNeasy Blood & Tissue Kit (Venlo, Netherlands) following the Alkek Center protocols. PCR amplification of the V4 region of the 16S rRNA gene was performed using primers 515F and 806R using 2 × 250-bp paired-end sequencing protocols. Raw sequence data are available on the National Center for Biotechnology Information under accession no. PRJNA1217883. The Harvard Chan Microbiome Analysis Core performed further analysis on raw sequences. Briefly, DADA2 v1.28 was used for grouping raw reads as amplicon sequence variants and assigning these taxonomically to the genus level using Silva v138.1. All samples passed a final reads cutoff of 10,000 with the exception of D1V, which had no raw reads. Genera were filtered keeping those with at least 0.01% relative abundance in at least 10% of samples for statistical analyses. Maaslin2 v1.14.1 was used to test associations of taxa with covariates. The vegan v2.6-4 R library was used for quantifying and comparing diversity indices. All analyses were performed in R version 4.3.0. α Diversity was evaluated via calculation of the Shannon diversity index; β diversity was indexed by Bray–Curtis dissimilarity and visualized as a principal coordinates analysis plot. Normality of operational taxonomic units (OTUs) by group was confirmed with the Shapiro–Wilk test and the Kolmogorov–Smirnov test. Comparisons of OTUs were performed with GraphPad Prism software.

Results

Safety and clinical outcomes.

All macaques tolerated 6 min of blue light exposure without any signs of behavioral distress or negative clinical impact. No consistent clinical improvement of the skin/implant margin was observed. Clinical improvement was subjectively assessed by decreased erythema and granulation tissue response. Some macaques demonstrated mild improvement, whereas other animals with more pronounced granulation tissue responses showed negligible changes to their margins. Representative images of macaques that showed improvement and showed no improvement following blue light treatment are shown in Figure 2. Macaque D, who had an area of eroded skin over a cortical screw, showed the most improvement, subjectively, out of the 8 animals. Macaque B also showed mild improvement at the implant margin. Other macaques, such as macaque L, displayed a more pronounced granulation tissue response at the implant margin and did not show clinical improvement during the study period (Figure 2).

Figure 2.

Representative photos for 3 macaques of the blue light exposure area over the study duration. The prestudy image for macaque B is unfortunately unavailable. Clinical improvement was most pronounced in macaque D as demonstrated by decreased erythema. Macaque L did not show clinical improvement as demonstrated by a sustained granulation tissue response. Due to the diversity of cranial implant morphologies for animals enrolled on study, we were unable to create a quantitative scoring system.

Culture results.

Culture results are shown in Table 2. All animals cultured S. aureus both before and after the study. The second most prevalent species was β-hemolytic Streptococcus, isolated from 7 macaques prestudy and 5 macaques poststudy. The third-most prevalent species cultured was Corynebacterium with C. ulcerans isolated from 6 macaques prestudy, 4 macaques poststudy, and C. renale was isolated from one macaque poststudy. Gram-negative organisms were less prevalent on culture with 2 E. coli isolates identified pre (macaques L and A) and isolated again from macaque L poststudy. Proteus mirabilis was also isolated from macaque L poststudy. Finally, Acinetobacter spp. were isolated from macaque D only poststudy. Two anaerobes were isolated poststudy (Prevotella in animal G and Bacteroides in animal T).

Table 2.

Aerobic and anaerobic culture results of implant margin cultures in 8 rhesus macaques

Prestudy cultures			Poststudy cultures
Isolate	N	Study IDs	Isolate	N	Study IDs
S. aureus	8	All	S. aureus	8	All
β-Hemolytic Streptococcus	7	S1, L1, D1, B1, P1, G1, T1	β-Hemolytic Streptococcus	5	D2, A2, B2, G2, T2
Corynebacterium ulcerans	6	S1, D1, B1, P1, G1, T1	Corynebacterium spp.	4 (3 C. ulcerans, 1 C. renale)	S2, P2, T2, B2 (C. renale)
E. coli	2	L1, A1	E. coli	1	L2
Klebsiella spp.	1	A1	Proteus mirabilis	1	L2
			Acinetobacter spp.	1	D2
			Prevotella spp.	1	G2
			Bacteroides spp.	1	S2

Antimicrobial susceptibility testing.

All isolates underwent antimicrobial susceptibility testing via MIC testing at a commercial veterinary diagnostic laboratory following their standard protocols. MICs were not reported for β-hemolytic Streptococcus, Corynebacterium spp., and anaerobes. For S. aureus isolates, all isolates were susceptible to penicillin, oxacillin/methicillin, amoxicillin, amoxicillin-clavulanic acid, imipenem, cephalexin, cefpodoxime, amikacin, gentamicin, azithromycin, erythromycin, mupirocin, chloramphenicol, clindamycin, trimethoprim-sulfamethoxazole (TMS), minocycline, and rifampin (Table 3). All isolates showed either intermediate (n = 4) or resistant (n = 12) susceptibility to cefovecin. Staphylococcus aureus isolates from animals A and D displayed resistance to fluroquinolones, including ciprofloxacin, enrofloxacin, and marbofloxacin. The prestudy isolate from animal A also showed resistance to doxycycline. Sporadic intermediate susceptibility results were noted to doxycycline for D1 and P1 isolates and to florfenicol for A1, B1, and L2 isolates. For the Gram-negative isolates, antimicrobial resistance was reported against amoxicillin, cephalexin, and doxycycline for E. coli, Klebsiella spp., and P. mirabilis (Table 4). All E. coli isolates were in addition resistant to amoxicillin-clavulanic acid, and the isolates from animal L were resistant to all third-generation cephalosporins tested. The Acinetobacter spp. isolated from animal D poststudy demonstrated susceptibility to imipenem, all third-generation cephalosporins tested, amikacin, gentamicin, and doxycycline, and resistance to fluoroquinolones and TMS.

Table 3.

Differing minimum inhibitory concentration antimicrobial susceptibilities for S. aureus

Study ID	PEN	OX/ME	CEFO	CIP	ENO	MBO	CLN	DX	FLOR
A1	0.12	0.5	2	R	>4	>4	0.25	R	8
A2	0.06	<0.25	1	R	>4	>4	0.25	<0.12	<4
B1	0.12	<0.25	2	S	<0.5	<0.5	0.25	<0.12	8
B2	0.06	<0.25	2	S	<0.5	<0.5	0.25	<0.12	<4
D1	0.12	<0.25	1	R	>4	>4	0.25	0.25	<4
D2	0.06	<0.25	1	R	>4	>4	0.25	<0.12	<4
G1	0.12	<0.25	2	S	<0.5	<0.5	0.25	<0.12	<4
G2	0.12	0.5	2	S	<0.5	<0.5	0.25	<0.12	<4
L1	<0.03	<0.25	1	S	<0.5	<0.5	<0.12	<0.12	<4
L2	0.12	0.5	2	S	<0.5	<0.5	0.25	<0.12	8
P1	0.12	0.5	2	S	<0.5	<0.5	0.25	0.25	<4
P2	0.12	0.5	2	S	<0.5	<0.5	0.25	<0.12	<4
S1	0.12	<0.25	2	S	<0.5	<0.5	0.25	<0.12	<4
S2	0.12	0.5	2	S	<0.5	<0.5	<0.12	<0.12	<4
T1	0.12	<0.25	2	S	<0.5	<0.5	0.25	<0.12	<4
T2	0.12	0.5	2	S	<0.5	<0.5	0.25	<0.12	<4

Prestudy isolates are designated with the number 1 following the animal ID and poststudy isolates are designated with the number 2 following the animal ID. Unshaded MICs indicate susceptibility, yellow-shaded MICs indicate intermediate susceptibility, and red-shaded MICs indicate resistant susceptibilities. When numeric MICs were not provided, S and R are used to designate susceptible and resistant results, respectively. Key: PEN, penicillin; OX/ME, oxacillin/methicillin; CEFO, cefovecin; CIP, ciprofloxacin; ENO, enrofloxacin; MBO, marbofloxacin; CLN, clindamycin; DX, doxycycline; FLOR, florfenicol. In addition, all isolates were susceptible to amoxicillin (>0.25), amoxicillin-clavulanic acid (>0.25), imipenem (S), cephalexin (>2), cefpodoxime (>2), amikacin (>2), gentamicin (>0.5), azithromycin (S), erythromycin (>0.25), mupirocin (>1), chloramphenicol (>4), TMS (>10), minocycline (>0.5), and rifampin (>0.03).

Table 4:

Minimum Inhibitory Concentration Antimicrobial Susceptibilities for Gram-negative isolates

Study ID	Isolate	AMX	AMC	IMP	CEPH	CEFO	CEFPO	CEFOT	CEFTZ	CEFTF
A1	E. coli	R	8	<0.25	>64	1	2	S	<0.12	<1
A1	Klebsiella spp.	R	4	<0.25	4	<0.5	<0.25	S	<0.12	<1
D2	Acinetobacter	N/A	N/A	S	N/A	S	S	S	S	N/A
L1 + L2	E. coli	>32	R	<0.25	>64	>8	>8	R	32	>8
L2	P. mirabilis	>32	N/A	N/A	>64	<0.5	<0.25	S	<0.12	<1
Study ID	Isolate	AMK	GM	CIP	ENO	MBO	DX	CHLOR	TMS
A1	E. coli	<2	<1	<0.06	<0.12	<0.5	1	<2	<20
A1	Klebsiella spp.	<2	<1	<0.06	<0.12	<0.5	1	<2	<20
D2	Acinetobacter	S	S	R	R	R	S	N/A	R
L1 + L2	E. coli	<2	<1	<0.06	<0.12	<0.5	1	8	<20
L2	P. mirabilis	<2	<1	<0.06	<0.12	<0.5	>16	4	<20

Pre-study isolates are designated with the number 1 following the animal ID and post-study isolates are designated with the number 2 following the animal ID. Unshaded MICs indicate susceptibility and red-shaded MICs indicate resistant susceptibilities. Key: amoxicillin (AMX), amoxicillin-clavulanic acid (AMC), imipenem (IMP), cephalexin (CEPH), cefovecin (CEFO), cefpodoxime (CEFPO), cefotaxime (CEFOT), ceftazidime (CEFTZ), ceftiofur (CEFTF), amikacin (AMK), gentamicin (GM), ciprofloxacin (CIP), enrofloxacin (ENO), marbofloxacin (MBO), doxycycline (DX), chloramphenicol (CHLOR), trimethoprim-sulfamethaxazole (TMS).

Microbiota analysis.

Swabs for microbiota analysis were collected from both the area exposed to blue light (designated as ‘B’ samples) and a separate area that was only treated with HOCl (designated as ‘V’ samples). Sample D1V unfortunately had no reads on 16S sequencing and thus DV samples were excluded from downstream paired analysis.

Taxonomy.

The mean number of OTUs was 75.2, 96.5, 83.9, and 81.0 for groups B1, B2, V1, and V2, respectively. There was no significant difference seen between the number of OTUs identified between treatment groups or timepoints (one-way ANOVA, P = 0.9018, F = 1,906, df = 3). The 20 most abundant genera are displayed in a bar chart (Figure 3). All animals had unique microbiota profiles with Staphylococcus, Streptococcus, and Corynebacterium OTUs representing the 3 top genera identified. Visually, these 3 genera were especially prominent for animals A, P, and T. Numerous anaerobic genera, including Fusobacterium, Porphyromonas, Peptoniphilus, Anaerococcus, Parvimonas, and Prevotella, were identified as contributors to the implant microbiome.

Figure 3.

Relative abundance of the top 20 bacterial genera identified on microbiota analysis for macaques treated with blue light + HOCl compared with HOCl only, before (PRE) and after study (POST). In addition to Staphylococcus, Streptococcus, and Corynebacterium, numerous anaerobic genera are identified, including Fusobacterium, Porphyromonas, Peptoniphilus, Anaerococcus, Parvimonas, and Prevotella. Every macaque demonstrated a unique microbiota profile that was not significantly different posttreatment.

α Diversity.

There was no significant difference seen between treatment groups or timepoints for α diversity. The median Shannon diversity index was 2.62, 2.58, 2.65, and 2.62 for groups B1, B2, V1, and V2, respectively (Figure 4).

Figure 4.

α Diversity as evaluated by the Shannon diversity index of the blue light + HOCl (B) and HOCl only (V) groups at timepoint 1 (prestudy) and timepoint 2 (poststudy). The box-and-whiskers plot displays the inclusive median. No significant difference between groups or timepoints was observed.

β Diversity.

The principal coordinates analysis plot of β diversity did not demonstrate clustering based on timepoint or treatment. Macaques A, T, and P were more closely clustered on the plot for timepoint 1 (Figure 5). We separately examined changes in the bacterial taxa we knew to be most prevalent from previous microbiota work, that is, Staphylococcus (Figure 6A), Streptococcus (Figure 6B), and Corynebacterium (Figure 6C), initially hypothesizing that we would see decreased relative abundance for catalase-positive Staphylococcus and Corynebacterium and potentially increased relative abundance for Streptococcus. No consistent patterns of increased or decreased relative abundance were observed for any of these taxa, regardless of treatment group.

Figure 5.

β Diversity as visualized as a principal coordinates analysis showing the top 2 axes by variance in microbiota communities explained. Circles represent the blue light + HOCl group, and the triangles represent the HOCl-only group, with each animal designated by a different color. Lines connect sample pairs indicating shifts in microbial community between timepoints 1 (prestudy) and 2 (poststudy). Points are generally most clustered by animal (color), indicating that the individual microbiota has a stronger effect on microbial community than the treatment or timepoint.

Figure 6.

Individual assessment of abundance changes for Staphylococcus (6A), Streptococcus (6B), and Corynebacterium (6C) by macaque, treatment and time point. Blue light + HOCl treatment is shown in reddish-orange and HOCl only treatment is shown in teal with pre-treatment abundance on the left (timepoint 1) and post-treatment abundance on the right (timepoint 2). No consistency was observed for changes in abundance for the top three species cultured nullifying our original hypothesis that blue light with HOCl treatment would decrease Staphylococcus and Corynebacterium colonization and increase Streptococcus.

Discussion

The goals of this study were to evaluate the effects of blue light therapy on the macaque cranial implant bacterial population as assessed via traditional culture methods and microbiota techniques. Culture-based methods have the advantage of selecting the most abundant and least fastidious bacterial species and in addition offer the ability to perform antimicrobial susceptibility testing to optimize treatment when clinically indicated. In this population, all S. aureus isolates were methicillin-sensitive in addition to being sensitive to multiple classes of first-line antimicrobials. Staphylococcus aureus isolates from macaques A and D did display resistance to fluoroquinolones; these macaques arrived at our institution from another research facility, and likely had previous exposure to fluoroquinolones compared with the practices at our institution. Another potential technique to evaluate the effects of blue light therapy would be to quantify the bacterial load at the implant margin before and after treatment. The gold standard approach for quantifying bacterial loads would involve biopsy of the implant margin, homogenization of a fixed mass of tissue followed by either plating serial dilutions on agar,²⁵ or quantitative PCR for the 16S rRNA gene.²⁶ Due to the known colonization of the implant margin site, biopsy would risk iatrogenic clinical infectious sequelae that would potentially affect the animals’ ability to participate in their primary neuroscience research tasks. Due to this risk factor, we did not pursue methodology to quantify total bacterial load.

Multiple previous studies have assessed common species comprising normal flora as well as bacterial colonization of cranial implants via aerobic culture and sensitivity and anaerobic culture.^12,13,27,28 The culture results reported from this study are consistent with the previous literature identifying Staphylococcus spp., Streptococcus spp., and Corynebacterium spp. as common Gram-positive species identified on aerobic cultures both associated with implants, and other sources in literature evaluating normal flora and antimicrobial resistance trends.^{9,10,12,27–30} Literature evaluating anaerobic colonization is more limited, but studies have also identified Fusobacterium spp.,^10,13 Peptostreptococcus spp.,¹³ Bacteroides fragilis,^10,13 Prevotella spp.,¹³ Parvimonas micra,¹⁰ and Porphyromonas spp.¹³ from cranial implants via culture-based methods.

The use of 16S rRNA gene sequencing to analyze the bacterial population of the cranial implant margin highlights the limitations of traditional bacterial culture methods. In this study, a total of 10 unique bacterial species were cultured from 8 macaques at 2 timepoints, of which only 2 bacterial species were identified on anaerobic culture: Prevotella and Bacteroides. In contrast, 16S sequencing identified a mean of 75 to 96 OTUs per group, which included taxa of low abundance and numerous anaerobic species (Figure 3). The cranial implant margin microbiota in 14 macaques was previously assessed at a different institution.¹⁰ The macaque skin margin microbiota results from the current study were slightly less diverse than previous work, which reported a mean Shannon diversity index of 4.71 (compared with a median of ∼2.6 for this study) and a mean of 126 observed OTUs (compared with a mean of 75 to 96 OTUs for this study).¹⁰ The macaques in the current study do show a similar taxonomic profile to previous findings, with Staphylococcus, Streptococcus, Corynebacterium, Fusobacterium, Porphyromonas, Peptoniphilus (previously classified as Peptostreptococcus), Anaerococcus, Parvimonas, Prevotella, Bacteroides, and Treponema genera noted as components of the skin margin. Genera identified in this population that were not identified in the previous study include Campylobacter and Acinetobacter, while the macaques from the previous study in addition had a higher prevalence of Proteus and Pseudomonas identified. Differences in the taxonomy charts between the 2 studies are likely impacted by several factors. First, macaques from the previous study all had CRC implants whereas macaques in the current study have only implanted microelectrode arrays. Culture swabs collected from CRCs in the previous study had a high prevalence of Proteus identified as compared with the current study where only one macaque cultured Proteus at one time point (L2). Communication between CRCs and the skin margin via microdefects in PMMA could provide a persistent source of Proteus colonization of the skin margin for macaques with CRCs. Different antimicrobial administration practices between institutions can also select for different populations. Proteus is an excellent biofilm former and often displays a multidrug-resistant phenotype and may outcompete less resistant and poor biofilm-forming species in environments with heavier exposure to antimicrobials. The single Proteus isolate cultured from L2 in this study displayed a susceptible phenotype to third-generation cephalosporins, TMS, fluoroquinolones, aminoglycosides, and chloramphenicol in contrast to Proteus isolates from the previous study that were resistant to TMS, fluoroquinolones, aminoglycosides, and chloramphenicol. As well as differences in the bacterial populations between the previous and current study, minor differences in 16S sequencing techniques and annotation may contribute to differences. The previous study used sequencing of the V4V5 region (515F and 926R primers) with taxonomic assignments made against the Greengenes database, whereas the current study only sequenced the V4 region (515F and 806R primers) and used Silva for taxonomic assignment. Expansions and improvements in databases and algorithms over the 6 y in between the studies can also contribute to differences in genera identified.

While use of the blue light was safely tolerated by all macaques, significant shifts in the microbiota or changes in α and β diversity were not observed following the 12, 6-min exposures administered over 4 wk using the device chosen. When selecting a device during study design, the primary goal was to optimize for practical implementation of this technology in NHPs. The most important feature of a blue light device was to have a flat, sealed surface that could be easily sanitized, given the nature of the polymicrobial communities and ABSL2 environment in which the device would be used. We also optimized for the shape and size of the device; some commercially available devices are designed for acne spot treatments or have array shapes that would not work (for example, shaped similar to a face mask or wand-shaped). The device chosen has an approximately triangular shape with a surface area of 2.5 in². Another optimization factor was the recommended exposure time and accurate timing of the exposure for consistency during the experiment. Some commercially available devices are recommended to be used for 10 to 20 min, which would not be practical or achievable for either our macaques or our busy investigators. The device chosen had an automatic 3-min timer, and manufacturer instructions were recommended for use no more than 3 times daily, per area. Another consideration for choosing a device was verifying that the device did not get too hot during the exposure period. This was determined by ensuring that the authors could tolerate the exposure period on their own skin without any discomfort. A different device initially acquired was immediately ruled out due to the heat of the bulb. Finally, we wanted to ensure that the availability and the price point of the device would allow for widespread adoption of this technology should it be effective; accordingly, the selected device was $99 at the time of purchase and readily ordered from a major online retailer.

The determinants of the bactericidal activity of blue light are related to the wavelength used, sensitivity of bacterial strains treated, the irradiance of the device, and the dose administered.^31,32 Because commercially available, over-the-counter blue light devices are classified as FDA-cleared, minimal information is available from manufacturers regarding actual power or dose, as only similarity to an existing approved device is required. On follow-up with the manufacturer, it was determined that the power of the device used in this study was 35 to 40 mW/cm², which equates to a dose of 14.4 J/cm² for a 6-min exposure. This is ∼50% of the target dose 30 to 54 J/cm² used in previous publications.³³ As previously stated, a 12-min exposure time was not practical for our macaques or our investigators. One potential option would be to perform the blue light exposure while the macaque was performing behavioral tasks. This would necessitate setting up a hands-free rig to position the device and in addition require a way to keep the device on longer than the 3-min timer allowed without having to repress the power button. It is noteworthy that commercially available blue light devices are packaged with goggles for eye protection due to the brightness of the LEDs. Because the macaques were head-fixed, we had the ability to position the device at the caudal or lateral aspect of the skin margin without concern for inadvertent ocular exposure. For use at the rostral/brow margin of cranial implants, ocular exposure may be a concern even with head-fixation. A final complication for the efficacy of blue light devices around cranial implants is achieving direct contact and minimizing the distance between the light surface and the implant margin. To minimize the interference of hair, we regularly trimmed the hair around the exposure site with mini clippers, being careful to avoid any resulting skin irritation from shaving the hair too closely. While we held the device in contact with each macaque’s implant margin, the 3D morphology of the PMMA did prove to provide some barriers to direct contact for some parts of the exposure area. A flat area of exposure will maximize the dose delivered; this may explain why macaque D’s erosion site had the most visually improved appearance among the study animals, as it was easier to achieve full contact of the device over his erosion site (Figure 2). During this study, we purposefully minimized mechanical debridement of crusts and granulation tissue at the implant margin to avoid irritation of the margin before blue light exposure. In retrospect, any crusting and granulation tissue response may have blocked penetration of the blue light and been an additional confounding factor for our negative results.

In addition to the 30 blue 415-nm LEDs, the device chosen for this project incorporated 10 red 630-nm LEDs into the light array. Phototherapy at 630 nm in humans is advertised as an antiaging and antiwrinkle treatment. Previous literature suggests that 630-nm light can stimulate fibroblasts and increase skin collagen,^34,35 which can improve aesthetic appearance and skin elasticity. It is possible that the 630-nm LEDs may play a role in contributing to the subjective improvement of the skin margin in some animals. However, further studies are needed to better form a direct comparison of 630-nm compared with 415-nm light phototherapy, and this was outside of the scope of the current study.

We purposefully designed this study to evaluate the blue light device in conjunction with HOCl treatment without a separate blue light–only control. Because standard implant margin maintenance at our institution commonly utilizes disinfectants (including dilute povidone-iodine and/or HOCl), we did not want to rely on an unproven technology as a sole method of sanitization. Because the purported antimicrobial efficacy of blue light is related to damage resulting from reactive oxygen species production, we were expecting that the antibacterial effects of a mild oxidizing agent would be enhanced following blue light exposure. Therefore, we determined that the most relevant comparison was to evaluate changes in the microbiota between the combination of blue light and HOCl and HOCl alone. By using each macaque as its internal control, we attempted to account for each macaque’s individual microbiota profile. This necessitated the underlying assumption that the microbiota is similar between 2 separate cranial implant margin sites within the same macaque. This assumption was reasonable for most of our study animals when comparing each macaque’s B and V prestudy samples on the taxonomy chart (Figure 3) and initial β diversity points on the principal coordinates analysis plot (Figure 5). Macaque D, unfortunately, had no sequencing reads from his V1 (prestudy HOCl only) sample. Macaque D’s HOCl only sampling location was around his headpost, as this animal had no external PMMA surrounding his headpost, and his headpost margin was clinically free of erythema and granulation tissue. It is unknown why the swab did not pick up normal skin flora, as it did for macaque A (who also had no PMMA around his headpost and a headpost margin free of erythema and granulation tissue).

In evaluating clinical efficacy, a limitation of this study was lack of an objective scoring system. The small number of animals enrolled and the diversity of the shapes and clinical appearances of their cranial implant complicated the creation of a relevant scoring system. Two animals (macaques A and D) had only headposts (without PMMA) and lacked any significant erythema, skin retraction, or granulation tissue response around their headpost margin at the beginning of the study. In contrast, macaques S and L had significant granulation tissue responses throughout the entire study at both the blue light and HOCl-only sites, which did not demonstrate visible change throughout the study period. If future studies are pursued, more consideration will be given to how to objectively score implant margins, especially if larger numbers of animals are enrolled. Another consideration for future studies will be to evaluate animals with more similar implant structures. This project was limited to animals with existing cranial implants; future work should also consider addressing use of blue light therapy in the immediate postoperative period to see if there are applications for minimizing bacterial colonization at the implant/tissue margin.

In conclusion, we found that the 415-nm blue light device used in this study was clinically safe and well tolerated by all 8 macaques; however, we did not see consistent or significant alterations in the cranial implant microbiota. The microbiota data from this study confirm our previous work, highlighting the contribution of anaerobic species to the implant/skin margin microbiota and emphasizing the limitations of culture-based methods to evaluate bacterial species present in these complex environments. Future work should aim to identify blue light devices that can provide a 30 to 54 J/cm² dose of 415-nm blue light to see decreases in OTUs, or if shifts in α and β diversity can be detected.

Conflict of Interest

The author has no conflicts of interest to declare.

Funding

This work was supported by Grants for Laboratory Animal Science (GLAS) from the American Association for Laboratory Animal Science.