8 Agonists for Targeted Delivery

Subscriber access provided by University of Massachusetts Amherst Libraries

Article

Hyaluronic Acid Conjugates of TLR7/8 Agonists for Targeted Delivery to Secondary Lymphoid Tissue Euna Yoo, Alex C. D. Salyer, Michael J. H. Brush, Yupeng Li, Kathryn L. Trautman, Nikunj M. Shukla, Ans De Beuckelaer, Stefan Lienenklaus, Kim Deswarte, Bart Lambrecht, Bruno G. De Geest, and Sunil A. David Bioconjugate Chem., Just Accepted Manuscript • DOI: 10.1021/acs.bioconjchem.8b00386 • Publication Date (Web): 03 Jul 2018 Downloaded from http://pubs.acs.org on July 4, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

TOC

Targeted Lymph Node Delivery Potent Adjuvantic Effect

Minimal Systemic Exposure Minimal Reactogenicity

ACS Paragon Plus Environment

Bioconjugate Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Page 2 of 46

Page 3 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Hyaluronic Acid Conjugates of TLR7/8 Agonists for Targeted Delivery to Secondary Lymphoid Tissue Euna Yoo,† Alex C. D. Salyer,†‡ Michael J. H. Brush,‡ Yupeng Li,‡ Kathryn L. Trautman,‡ Nikunj M. Shukla,† Ans De Beuckelaer,§ Stefan Lienenklaus,¶ Kim Deswarte,§ Bart N. Lambrecht,§ Bruno G. De Geest,§ and Sunil A. David‡*

† Department of Medicinal Chemistry,

University of Kansas, Lawrence, KS 66045, USA. ‡ Department of Medicinal Chemistry,

University of Minnesota, Minneapolis, MN 55455, USA. § Department of Pharmaceutics and Center for Inflammation Research, Ghent University, 9000 Ghent, Belgium. ¶ Institute for Laboratory Animal Science, Hannover Medical School, 30625 Hannover, Germany.

Running Title:

Lymph node delivery of hyaluronic acid conjugated adjuvants

Keywords:

TLR7, TLR8, TLR7/8 agonists, Hyaluronic acid, Vaccine adjuvants, Innate immunity, Lymph node, Targeted Delivery.

1 ACS Paragon Plus Environment


Page 4 of 46

Abstract

Immunogens carried in lymphatic fluid drain via afferent vessels into regional lymph nodes and facilitate the efficient induction of appropriate immune responses. The lymphatic system possesses receptors recognizing hyaluronic acid (HA). Covalent conjugates of smallmolecule TLR7/8 agonists with HA are entirely devoid of immunostimulatory activity in vitro. In murine models of immunization, however, such conjugates traffic to lymph nodes, where they are ‘unmasked’, releasing the small molecule TLR7/8 agonist from the carrier polysaccharide. The resulting focal immunostimulation is manifested in potent adjuvantic effects with negligible systemic exposure. The efficient delivery of immunogens has been a major challenge in the development of subunit vaccines, and enhancing targeted delivery of immunogens to secondary lymphoid organs might be a promising approach for improving vaccine efficacy, as well as safety.

(126 words).


Page 5 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction

The lymphatic system provides for unidirectional transport of transudative interstitial fluid and proteins exiting microcirculation due to hydrostatic pressure, and returning them back to the intravascular compartment. Lymphatic fluid drains via afferent lymphatic vessels into regional lymph nodes, which are highly organized secondary lymphoid tissues specialized in immune surveillance of the contents of lymph.1 Immune cells located in specialized zones within lymph nodes not only respond to antigens arriving from distal sites of infection, but also receive and orchestrate appropriate immune responses to migrating antigen presenting cells (APCs) bearing antigenic epitopes. This is of particular relevance in immunization, wherein peripheral APCs such as dendritic cells (DCs) and macrophages capture antigens from the injection site, and then migrate into the lymphoid tissues to trigger downstream Tand B-lymphocyte activation as well as memory cell differentiation.2 The lymph nodes also contain a large number of resident APCs, which can actively process antigens and serve as a major reservoir for long-lived memory B cells and central memory T cells, therefore playing a crucial role in generating long-term immunological memory.3, 4

The flow of interstitial fluid also brings fragments of extracellular matrix (ECM) macromolecules into lymph.5, 6 Important among the constituents of ECM is hyaluronic acid (hyaluronan, HA), a linear glycosaminoglycan polymer with a molecular weight that can reach 107 Daltons, and composed of repeating polymeric disaccharides of D-glucuronic acid (GlcUA) and N-acetyl glucosamine (GlcNAc) linked by β-1,4 and β-1,3 glycosidic bonds (Figure 1).7 Depending on the molecular size, HA can interact with several cognate HArecognizing molecules8 important among which are CD44, a single chain, transmembrane 3 ACS Paragon Plus Environment


Page 6 of 46

glycoprotein expressed on leukocytes that traffic through the lymphatics,9 as well as lymphatic vessel endothelial hyaluronan receptor-1 (LYVE-1), which is expressed almost exclusively on lymphatic endothelium.10 CD44–HA interaction is known to be involved in a variety

of

cellular

functions,

including

cell–cell

interactions,

receptor-mediated

internalization/degradation of HA, and cell migration. LYVE-1 participates in HA-mediated leukocyte trafficking, adhesion, and transmigration.6

Figure 1. Chemical structure of hyaluronic acid.

The efficient delivery of antigen/adjuvant has been a major challenge in the development of subunit vaccines, and enhancing vaccine delivery to secondary lymphoid organs might be a promising approach for improving vaccine efficacy.2, 11, 12 Several studies have addressed enhanced or targeted delivery of antigens to secondary lymphoid tissue, including the use of depot-forming adjuvants,13 nanoparticulate carriers that are preferentially internalized by APCs,14-16 or intralymphatic immunization,3 but strategies that could use well-defined molecular conjugates would be more attractive. We envisioned that we could take advantage of the characteristics of HA, a natural biopolymer with an excellent clinical track record17-19 — biodegradability, biocompatibility, high potential loading and, importantly, its known propensity for trafficking to lymph nodes.20,

21

We hypothesized that covalently coupled 4


Page 7 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


conjugates of HA bearing TLR7/TLR8 agonists, which we have previously demonstrated to induce prominent Th1-biased responses22-25 could facilitate targeted delivery of these adjuvants to secondary lymphoid tissue while also minimizing systemic exposure of these small molecules with molecular properties that portend a large volume of distribution.

We show that covalent conjugates of small-molecule TLR7/8 agonists with HA are entirely devoid of immunostimulatory activity in vitro. In murine models of immunization, however, such conjugates traffic to lymph nodes, where they are ‘unmasked’ resulting in focused immunostimulation and potent adjuvantic effects with negligible systemic exposure.



Page 8 of 46

Results and Discussion

For proof-of-concept studies, we selected four lead adjuvants undergoing preclinical evaluation as vaccine adjuvants in our laboratory; these included the pure TLR7 agonist 1,26 the pure TLR8 agonist 2,23 and the dual TLR7/8 agonists 3 and 4 (Figure 2).25, 27, 28

Figure 2. Chemical structures of TLR7 and TLR8 agonists used for hyaluronic acid conjugation.

Each of these TLR-active compounds bears an aliphatic primary amine (Figure 2) that could be utilized in a straightforward manner for direct coupling with carboxylic acid groups on HA. Of the numerous methods that have been reported for HA conjugation,7 we avoided methods entailing harsh reaction conditions such as strongly alkaline or acidic pH, prolonged heating, or conditions calling for ultra-sound or microwave irradiation, which are known to potentially induce significant HA degradation, and elected to use the recently-reported 26 ACS Paragon Plus Environment

Page 9 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


chloro-4,6-dimethoxy-1,3,5-triazine (CDMT)-activated amidation strategy, since it allows for a highly controlled substitution under aqueous/mixed-solvent conditions, at room temperature, and at neutral pH (Scheme 1).29, 30

Scheme 1. Synthesis of HA Conjugates.

All reactants were soluble in a mixture of water and acetonitrile (3:2) and the pH values of the reaction mixtures after addition of all reactants were about 7. After overnight stirring at room temperature, the reaction mixtures were treated with Dowex ion exchange resin (hydrogen form) before dialysis to remove the morpholinium species, as well as unreacted amines which could otherwise act as counter-ions in their protonated form toward nonmodified carboxylic acid groups of HA. Exhaustive dialysis against 0.1 M NaCl also helped eliminate electrostatically bound amines from anionic polymers. Consistent with our prior 7 ACS Paragon Plus Environment


experience with chemical transformations on these small molecules,25,

Page 10 of 46

28, 31

the benzylic

amine, and not the amidine amine, participates in the formation of the amide bond, and reactions with analogues that do not possess aliphatic primary amines did not yield detectable substitution (data not shown). Thus, the coupling of the individual TLR7-, TLR8-, and TLR7/8-agonistic compounds 1-4 via the aliphatic primary amine with the carboxylates on HA yielded the conjugates HA1-HA4, respectively. A comparison of 1H NMR spectra of the purified conjugates HA1-HA4, with that of native hyaluronic acid and of the small molecules 1-4 (Figure 3) showed that the poorly-resolved broad multiplets between c.a. 3.2 – 4.0 ppm correspond to the resonances of the glycosidic protons of HA; a broad peak at 4.5 ppm could be assigned to the two anomeric protons attached to the carbons adjacent to two oxygen atoms, and the CH3 protons of the N-acetyl group of HA could be readily identified at 1.95 ppm (Figure 3). New peaks associated with aliphatic (butyl or pentyl chain) protons at 0.8 – 3.0 ppm and aromatic protons at 7.0 – 8.5 ppm, both of which are diagnostic of the conjugated small molecules, were observed.


Page 11 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. 1H NMR spectra in D2O of unconjugated hyaluronic acid, small molecules 1-4 and the conjugates HA1-HA4.The ω-protons of the C2-butyl/pentyl groups in the small molecules and in the corresponding conjugates are shown shaded in red. The methyl protons of the GlcNAc moieties of HA are shaded green.



Page 12 of 46

The degrees of substitution (DS) were determined by comparing integrated signals of the ωCH3 resonance of the butyl/pentyl side chain of the small molecules (triplet, 0.86 ppm); this resonance could be clearly resolved from that of the CH3 signal arising from the Nacetylglucosamine moiety (singlet, 1.95 ppm, Figure 3) of native HA. As shown in Table 1, the DS ranged from 29 to 39%. The slight difference in the observed DS could be attributable to differential steric accessibility of the amines and/or solubility.

Table 1. Properties of the conjugates. DS, degree of substitution; SM, small molecule. Conjugate TLR agonist (SM) No.

Yield (%)

DS (%)

SM Content (w/w %)

HA1

1

75

29

21

HA2

2

75

35

10

HA3

3

71

39

31

HA4

4

74

36

21

While the ratios of peak integrals by 1H NMR spectroscopy showed high batch-to-batch repeatability and consistency, we were mindful of the broad signals due to highly viscous sample solutions, as well as potential sample microheterogeneity, and we therefore sought to quantify the ‘loading’ and characterize purity of the conjugates using alternate methods. When subjected to size exclusion chromatography (SEC) analyses, HA, as expected, has no significant absorption in the UV, whereas the conjugates, displaying characteristic absorption-spectral profiles, were clearly observed to elute with a retention time of 50 min, corresponding to the void volume of the column (Figure S1A). Reverse-phase HPLC analyses were also carried out to ensure the purity of conjugates. With a polymeric reverse10 ACS Paragon Plus Environment

Page 13 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


phase PRP-3 column (poly(styrene-divinylbenzene), 300 Å pore size), conjugate HA3, for example, eluted at 32 min, while that of its corresponding small molecule 3 appeared at a retention time of 14 min. Again, the spectroscopic signatures of the small molecules enabled unambiguous identification of the conjugates and, importantly, verified that no unconjugated small molecules were detected in the conjugate samples (Figure S1B).

In vitro enzymatic degradation of the conjugates was performed using hyaluronidase (HAase) from bovine testes (400-1000 units/mg), and the fragments analyzed by LC-MS to identify oligosaccharides of HA covalently adducted with the small molecules. In all four conjugate digests, fragments corresponding to di-, tetra-, or decasaccharides with one small molecule conjugated, as well as dodecasaccharides fragments bearing two small molecules were identified (Table 2, Figure S2).

Table 2. Major oligosaccharide species found by LC-MS analysis after in vitro hyaluronidase enzymatic degradation of hyaluronic acid conjugates.

Conjugate No.

RT (min)

Conjugate HA1

7.262

[GlcUA-GlcNAc] –1

7.927

[GlcUA-GlcNAc] –1

8.885

[GlcUA-GlcNAc] –[1]

6.966

[GlcUA-GlcNAc] –2

7.759


8.505

[GlcUA-GlcNAc] –2

Conjugate HA2

Species

5

Conjugate No.

RT (min)

Conjugate HA3

6.234

[GlcUA-GlcNAc] –3

6.965


7.549

[GlcUA-GlcNAc] –3

6.059

[GlcUA-GlcNAc] –4

6.830


7.247

[GlcUA-GlcNAc] –4

2

6

2

Conjugate HA4

2

6

2

5

Species

2

6

2

5

2

6

2

5



Next, the conjugates were screened in NF-κB reporter gene assays specific for human TLR7 and TLR8. Whereas all small molecule TLR-7 and/or -8 agonists (1-4) showed, as expected, dose-dependent agonistic activities in cognate reporter gene-based assays,25, 27, 28, 32-35 we were surprised to find that none of the HA conjugates displayed any detectable NF-κB induction up to concentrations of 50 µg/mL in both TLR7 and TLR8 agonism assays; hyaluronic acid (used as a control) was also verified to be quiescent in these assays (Figure 4). These results were unexpected, given that derivatization of the amine in dual TLR7/8 agonists such as 3 to yield a variety of conjugates,25, 28, 31, 35 including a large dendrimeric species35 had previously resulted in loss of TLR8 activity, but with preservation of TLR7agonistic activity.

Figure 4. Direct comparison of the activity of the unconjugated small molecules (1-4) and their corresponding HA conjugates (HA1-HA4) in human TLR7/TLR8-specific reporter gene assays. Xaxes denote gravimetric concentrations for both small molecules and conjugates. The molecular weights of the small molecules 1-4 range between 369-450 Da, while those of the conjugates HA1HA4 are estimated to be > 2 X 106 Da. 1.2

1.5

hTLR8

hTLR7

Small molecules: 1.2

Relative NF-kB Induction

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 46

0.9

0.9

1: TLR7 agonist 2: TLR8 agonist 3: TLR7/8 agonist 4: TLR7/8 agonist HA conjugates:

0.6

0.6

0.3

Controls:

HA1 HA2 HA3 HA4

HA Neg. Ctrl.

0.3

0.0

0.0

10

-2

10

-1

10

0

10

1

10

-2

10

-1

10

0

10

1

Concentration of small molecule or conjugate (µg/mL) 12 ACS Paragon Plus Environment

Page 15 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


We examined the cytokine/chemokine-inducing properties of the conjugates in human PBMCs. Consistent with the lack of activity in reporter gene assays described above, the conjugates were entirely devoid of proinflammatory cytokine induction (represented by IL-1β, IL-6, IL-8, and TNF-α in Figure 5). A plausible interpretation is that the hyaluronic acid conjugates become “silent”, presumably because the polyanionic character of the conjugates precludes significant internalization and access to the endolysosomal compartment. Indeed, no detectable uptake of HA conjugated with rhodamine (used as a model system) in either whole human blood or in human peripheral blood was detected using flow cytometry (Figures S3 and S4, respectively) and fluorescence microscopy (data not shown). The premise that the HA conjugates undergo negligible internalization in PBMCs is supported also by the observation that these conjugates not only do not induce proinflammatory cytokines, but are also indistinguishable from negative controls in the induction of CC and CXC chemokines (data not shown).

These unexpected in vitro observations presented two possible outcomes in vivo: (a) the HA conjugates on account of being ‘silent’ in vitro would be entirely without adjuvantic effect in animal models; (b) given the extensive metabolism and turnover of HA in tissue,36-38 degradative mechanisms operational in tissue (and not observed in isolated PBMCs) could result in the hydrolysis of the amide bond-linked small molecules in vivo, resulting in observable effect(s).



Figure 5. Dose-response profiles of cytokine induction in human PBMCs by the selected conjugates (HA3 and HA4) and small-molecule TLR7/8 agonist comparators (3 and 4). Responses to 3 and 4 were essentially indistinguishable. X-axes denote gravimetric concentrations for both small molecules and conjugates. The molecular weight of 3 and 4 (regioisomers, HCl salts) is 395 Da, while the estimated molecular weights of the conjugates HA3 and HA4 are > 2 X 106 Da. 5000 600

400

IL-1β

IL-6 4000

Compound 3 Conjugate HA3 Conjugate HA4 Unconjug. HA Neg.Ctrl.

3000

2000

Analyte Concentration (pg/mL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 46

200

1000

0

0

1500

IL-8

TNF-α

6000

1000 4000

500 2000

0

0

1

10

1

10

Concentration of small molecule or conjugate (µg/mL)


Page 17 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In order to test these possibilities, we examined these conjugates for potential adjuvantic properties in New Zealand White rabbits. Animals (n = 4 per cohort) were immunized with a standardized protocol with Diphtheria toxoid (CRM197)39 as the antigen at 10 µg/dose and the hyaluronic acid conjugates at 100 µg/dose, formulated under strictly excipient-free conditions in sterile, pyrogen-free saline. Rabbits were pre-bled on Day 0 for estimation of preimmune titers, and then primed with the first dose of vaccine. Animals were boosted on Days 15 and 28, and bled on Days 25 (for Immune-1 sera) and 38 (for Immune-2 sera), respectively.

As shown in Figure 6, all of the conjugates displayed adjuvantic activity, with conjugate HA4 being extraordinarily potent, evoking high antigen-specific IgG titers even after a single boost, resulting in titers significantly higher than that obtained using the corresponding unconjugated small molecule TLR7/8 agonist 3 (used as a comparator, Figure 6). Considering the very small effective dose of the TLR agonist in the conjugate, the strong adjuvantic activity of the conjugate after a single boost, and the complete lack of proinflammatory cytokine induction in ex vivo models, as described earlier, presaged a highly effective and safe vaccine adjuvant with negligible local or systemic reactogenicity.40



Figure 6. Adjuvantic activities of HA conjugates after a single boost. Antigen: Diphtheria toxoid (CRM197), 10 µg/dose; Adjuvant: 100 µg/dose in NZ rabbits (n=4/cohort); Formulation: excipientfree saline. A. Box-and-whisker plots of antigen-specific titers in HA conjugate-adjuvanted rabbits, compared with hyaluronic acid (control), and compound 3 (comparator). Plots depict mean, median, 1st and 99th percentiles. B. Quantitation of antibody avidity in chaotropic ELISAs with NaSCN concentration eliciting 50% attenuation of absorbance.

4

A

B

2.0

3

Absorbance at 50% attenuation

Log10(Immune 1/Preimmune IgG Titers)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 46

2

1

0

Conjugate HA4

1.5

Conjugate HA3

1.0

Cmpd 3 0.5

0.0

HA (Control)

HA1

HA2

HA3

HA4

Cmpd

3

(Comparator)

0.0

0.5

1.0

1.5

2.0

NaSCN Concentration (M) eliciting 50% Attenuation of Absorbance

While measurements of serum antibody titers employing an immunoassay such as ELISA serve to quantify antigen-specific antibody concentrations, antibody ‘avidity’ (functional affinity; the sum of affinities of multiple antigen-binding sites in polyclonal antibodies simultaneously interacting with their cognate antigenic epitopes) is an important characteristic of protective immune responses.41 In order to also assess the quality of antibody, the strength of the interaction between CRM197 and antigen-specific IgG antibodies obtained 10 days after a single boost (Immune-1) was determined using the chaotropic ELISA.42, 43 Replicate wells containing antibody bound to antigens were exposed to increasing concentrations of the chaotropic thiocyanate ion that disrupts antibody-antigen 16 ACS Paragon Plus Environment

Page 19 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


complexes and the concentration of sodium thiocyanate (NaSCN) that decreases antibody binding by 50% was used as the measurement of avidity. As shown in Figure 6B, the highest avidity was found in the conjugate HA4-adjuvanted sera with 50% attenuation of absorbance elicited at 1.0 M NaSCN. This result indicates high quality, affinity-matured IgG elicited by the hyaluronic acid conjugate. These in vivo studies unambiguously demonstrated—rather counter-intuitively, however— that despite the fact that the hyaluronic acid conjugates are entirely bereft of TLR-agonistic activities in vitro, both in respect to TLR recognition as probed by the reporter gene assays, as well as in their ability to induce cytokines in secondary screens, possess very pronounced adjuvantic properties in a standardized animal model. Indeed, conjugate HA4 has proven to be the best adjuvant that we have characterized to date, evoking, with a single boost regimen, affinity-matured, high-avidity immunoglobulins whose titers rival and surpass all of the best-in-class small molecule adjuvants that have been evaluated by us during the past decade. Importantly, its inertness in vitro presages very low systemic exposure and, consequently, a high margin of safety.

We sought to explore the disposition of conjugate HA4 in tissue, and whether significant trafficking to secondary lymphoid organs occurred, as we have recently observed with pHdegradable imidazoquinoline-ligated nanogels.25 Our first objective was to examine whether imidazoquinoline conjugates of HA would be transported to draining lymph nodes. We addressed this question by synthesizing an alkyne-bearing TLR7/8-active analogue of 4 (1(3-(aminomethyl)-5-methyl-4-(prop-2-yn-1-yloxy)benzyl)-2-butyl-1H-imidazo[4,5-c]quinolin-4amine, 9, Scheme S1), conjugating the imidazoquinoline to HA, and then covalently coupling a highly water-soluble, fluorescent probe with a high quantum yield (sulfo-cyanine3 azide) 17 ACS Paragon Plus Environment


Page 20 of 46

via a Cu-catalyzed ‘Click’ reaction,44 using the strategy outlined in Scheme S1. Subcutaneous or footpad injections of this fluorescent HA-IMDQ-Cy3 ternary conjugate allowed the direct visualization of trafficking. Fluorescent microscopy on the draining (ipsilateral popliteal and inguinal) lymph nodes showed the appearance of the HA-IMDQ-dye conjugate in 4 h, localizing, as expected for macromolecular cargo in afferent lymph, first to the subcapsular sinus, and then to the subjacent cortical sinuses (Figure 7).


Page 21 of 46

Figure 7. Fluorescence microscopy of representative draining (ipsilateral popliteal and/or inguinal) lymph nodes harvested at 4 h (panel A) and 24 h (panel B) from mice injected in the footpad with 10 µg of fluorescent HA-IMDQCy3 ternary conjugate. Red channel corresponds to the HA-IMDQ-Cy3; green channel represents CD11c+ dendritic cell subsets, and blue channel is B220+ B lymphocytes. Time-dependent influx of Cy3+ lymphocytes into the stroma of the draining lymph node (panels C, D). Lymphocytes were gated based on their FSC and SSC properties. Dead cells were excluded using a live-dead stain and doublets/multiplets were excluded.

A

B

800

C

600

400

200

D

5

A

4x10

Cy3+ Lymphocyte Counts

(Cy3 Channel, Arbitrary Units)

Lymphocyte-associated Median Fluor. Intensity

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


5

3x10

5

2x10

5

1x10

0

0

20

40

Time (h)

0

20

40



Page 22 of 46

We reasoned that the complete lack of TLR-agonistic activities of the HA conjugates in vitro on the one hand, and very pronounced adjuvantic properties in vivo, on the other, likely signified release of the imidazoquinoline moiety, either as small oligosaccharide fragments bearing the covalently bound TLR7/8 agonist, or hydrolysis of the amide bond, and subsequent release of the small molecule from the hyaluronic acid polymer via the action of lymph node hydrolases.45,

46

Our second objective, accordingly, was to directly visualize

innate immune activation in the lymph node. We utilized whole-body imaging in IFN-β reporter mice in which a firefly luciferase encoding sequence had been placed under the control of the IFN-β promoter,47 allowing for precise spatiotemporal analyses of IFN-β induction in response to TLR7/8 engagement within lymphatic tissue. We have recently shown that this model is useful not only in examining tissue disposition of the adjuvant, but also in assessing systemic exposure.25

Mice were injected in either hock with 10 µg of either the unconjugated small-molecule imidazoquinoline 4, or an equivalent dose of HA4, followed by full-body luminescence imaging of the luciferase signal at 4, 8, 24, 48, 72, and 96 h post-injection. The unconjugated small molecule 4 rapidly induced a high-magnitude, widespread systemic IFN-β response, either when administered alone, or as an admixture with hyaluronic acid, likely attributable to the rapid diffusion of the small molecule into systemic circulation (Figure 8B, 8C).


Page 23 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In stark contrast, HA4 induced significantly weaker luciferase signals localizing to the draining (inguinal) lymph nodes; mice injected within one hock showed signals in the ipsilateral, and not contralateral lymph node (e.g., mouse 2 in Figure 8A) throughout the experiment, pointing to possible sequestration of the conjugate in the lymph node. Weak signals were also observed at the site of injection for 96 h (the experimental endpoint) pointing to a depot effect. Interestingly, mice injected with HA4 pre-treated with hyaluronidase showed pronounced injection site signals without any appreciable signals arising in the lymph nodes (Figure 8D).

Figure 8. Luminescence imaging in IFN-β reporter mice. Animals were injected in the hock with conjugate HA4 (A), the unconjugated imidazoquinoline 4 (B), a mixture of imidazoquinoline 4 and hyaluronic acid (C), or conjugate HA4 digested with hyaluronidase (D). Red arrows in panel A denote the injection site (hock); high resolution X-ray co-registration shows signals from the draining lymph nodes (inguinal, yellow arrows) in animals injected with conjugate HA4 (panel A). Animals receiving hyaluronidase-digested HA4 show signals localizing to the injection site, but not to the lymph nodes (panel D). Luminescence intensity (photons/sec/cm2/sr) is represented in false colors. High-resolution X-ray co-registration signals in panel A were obtained with a Bruker In-Vivo Xtreme II also show signals localizing to the submandibular glands which was verified to be present in control animals injected with luciferin, pointing to possible salivary clearance of luciferin. (E) Temporal evolution of whole-body luminescence signals.



A

B

C

D

7

3x10

Total Flux (Photons/sec)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 46

E

Conjugate HA4: M1 M2 M3 M4

HA+Compound 4: M9 M10 M12 HA-ase Digest of HA4: M6 M7 M8

7

2x10

Compound 4: M15 M16 M13 M14

HA (Control): 1 M17 A M18 a M19 M20

7

1x10

a 0 A 1 0

2 b B

C c 3

D 4 d

E 5 e

g G 7

F 6f

30

60

h H 8

90

Time (Hours)


Page 25 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


It was of particular interest to examine if signals emanating in the draining lymph nodes of mice receiving conjugate HA4 were a consequence of degradation of oligosaccharide fragments, or of the hydrolysis of the amide bond-linked imidazoquinoline. Hyaluronidase digests of HA4 were inactive in in vitro TLR7/8 reporter assays (Figure S5), pointing strongly to the latter possibility. We interrogated lymph node homogenates by high-sensitivity triplequadrupole mass spectrometry for the presence of the imidazoquinoline 4 (M+H+: 360.2 Da), as well as the 4-D-GlcUA (536.2 amu) and 4-D-GlcUA-GlcNAc (739.3 amu) fragments. The imidazoquinoline 4 could be easily identified (Figure 9); trace quantities of 4-D-GlcUAGlcNAc (738.3 amu) disaccharide fragment (distinguished by characteristic neutral loss of the terminal GlcNAc residue) were also identified, but not the 4-D-GlcUA monosaccharide fragment (Figure 9), providing direct evidence for the hydrolytic liberation of 4 in the lymph node.

Taking into consideration the known pathways of HA catabolism,36-38 as well as the observation that oligosaccharide fragments of the conjugate HA4 are quiescent in vitro, we speculate that HA4 is degraded by conventional homeostatic degradation of the highmolecular-weight conjugate by the tissue hyaluronidases Hyal-1 and -2,48 generating small oligomers that are subsequently broken down to disaccharide fragments in lysosomes,37 the hydrolase- and amidase-rich environment of which ultimately releases the imidazoquinoline 4. We do not yet know if the localization of HA4 in the lymph node is CD44-dependent. It is of relevance in this context to note that HAase digests of HA4 show marked depot effect with no detectable lymph node trafficking (Figure 8D), emphasizing the importance of the size of the conjugate, as well as pointing to the potential involvement of CD44.49-51



Figure 9. A. The small molecule 4 (M+H+: 360.2 Da, QQQ mass spectrometry) was detected in lymph node homogenates harvested 48 h following hock injection with HA4. B. Characterization of authentic 4-DGlcUA-GlcNAc (739.3 amu) fragment from HAase digest of HA4, showing neutral loss of the terminal GlcNAc. C. Detection of 4-D-GlcUA-GlcNAc fragment in lymph node homogenates harvested 48 h following hock injection with HA4. Integrated Area: 360.2 Daltons

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 46

5

4x10

A

5

3x10

LN Homogenate: Naive Mice LN Homogenate: HA4 (48 h)

5

2x10

5

1x10

0 1

2

3

4

Mouse Number

B

C


Page 27 of 46

The potent adjuvantic property of HA4 was reconfirmed in a large (n=16) rabbit immunization experiment, which not only permitted a careful evaluation of dose-sparing effects of antigen (Figure S6), but also allowed the examination of potential immune responses to hyaluronic acid. As observed in the original screen (Figure 6), robust antiCRM197 IgG titers were elicited after a single boost (Figure 10). Statistically insignificant, low titers of immunoglobulins directed against hyaluronic acid were observed in one animal, and were not detected in 15 animals, allaying concerns that derivatization of an endogeneous biopolymer may serve as a neoantigen.

Figure 10. Rapid induction of robust anti-CRM197 titers in a large (n=16) cohort of rabbits after a single boost with 10 µg of CRM197, adjuvanted with 100 µg HA4, and formulated in excipient-free saline. A. AntiCRM197 titers. Titers in Immune-1 and Immune-2 samples are statistically significant (p

8 Agonists for Targeted Delivery

Recommend Documents