Systemically Administered Plant Recombinant Holo-Intrinsic Factor Targets the Liver and is not Affected by Endogenous B12 levels

Precision targeting imaging agents and/or treatment agents to select cells or organs in the body remains a significant need and is an area of intense research. It has been hypothesized that the vitamin B12 (B12) dietary pathway, or components thereof, may be exploitable in this area. The question of whether gastric Intrinsic factor (IF), critical for B12 absorption in the GI tract via the cubilin receptor, could be used as a targeting moiety for the cubilin receptor systemically, has not been investigated. Cubilin is the only known receptor for holo-IF and is found primarily in the kidney and ear (outside of the ileum of the GI) offering significant scope for specific targeting. We utilized plant derived human gastric IF in fluorescent cell and PET based in vivo imaging and biodistribution studies and demonstrated that plant derived IF primarily targets the liver, likely a consequence of the unique glycosylation profile of the IF, and is not affected by endogenous B12 levels.

the holo-TC receptor, CD320. Upon internalization, TC is degraded and B12 is transported from the lysosome for cellular use.
Herein, we sought to investigate the effects of systemic administration of B12 conjugates pre-bound to recombinant human gastric IF. The first outcome postulated would be that IF pre-binding would facilitate targeting the only known holo-IF receptor, CUBN, located in the ileum in the enterocyte, as described for dietary uptake, but also in the proximal tubules (PT) of the kidney, where it now partners with megalin and plays a role in reabsorption of such ligands as albumin, transferrin, vitamin D binding protein, apolipoprotein AI, amongst others 24 . We also postulated that IF pre-binding would prevent TC binding and hence would not be affected by endogenous B12 levels, a long-time concern in the field given binding to TC results in significant back-ground across tissues and offers the possibility of causing a loss of B12 cellular delivery (TC dependent) 25 . Expression of CUBN elsewhere is limited, including the human inner ear 26 and yolk sac 27 .
Before beginning such work, it was necessary to ensure access to IF that (1) was available commercially on a large-scale (i.e. 30-50 mg quantities) necessary to conduct, and ultimately translate the work, and (2) that it was in the apo-(i.e. no pre-bound B12) form, to allow binding of the desired B12-conjugates, which in this case are radio-probes of 89 Zirconium-B12 ([ 89 Zr]-B12), vide infra 28 . To achieve this, the only available source meeting our criteria was human recombinant IF (IF) produced in the plant Arabidopsis thaliana 29 . Expression in plants produces apo-IF, given plants are a rare organism that do not use B12, minimizing holo-IF production in situ. Questions to be explored with A. thaliana produced IF included the glycosylation profile of such a protein and the effects of such glycosylation on receptor targeting in vivo, as noted above, and whether this profile negated, complemented or refocused the CUBN targeting hypothesis noted above. The glycosylation of IF was examined by GC-MS (Table 1 and Supplementary Fig. S4). The sugars identified were α(1-3)-fucose, xylose, mannose and n-acetylglucosamine in the ratios 0.   B12-Cy5 to show functionalization of the IF-B12 complex in vitro (Fig. 2). Results showed no association of B12-Cy5 alone, and significant association of IF-B12-Cy5 at 37 °C. Reduction in binding (or internalization) of IF-B12-Cy5 at 4 °C supported a receptor mediated internalization. No association/binding was observed in Chinese hamster ovary (CHO) cells (CUBN and CD206 free cells; CD320 (TC receptor; XM_027442179.1)+; Figure S7) or in ASGPR positive ( Figure S6) HepG2 cells ( Figure S8). Then, we investigated uptake in J774.A1 macrophage cells (CUBN-and CD206+) 30 , which again showed no binding of B12-Cy5 alone, and binding of IF-B12-Cy5 at 37 °C. Adding mannan (2 mg/mL), 45 minutes prior to, and concomitant with IF-B12-Cy5 incubation, reduced IF-B12-CY5 uptake (Fig. 2) Fig. 3 and Table S1 there was significant liver uptake at 5 h, which did not change over the subsequent 24 h. Experiments were duplicated in mice on a B12 deplete diet for 21 days. For IF-[ 89 Zr]-B12 the highest uptake was seen in the  liver and kidneys and did not look significantly different than mice on replete diets (Fig. 3). However, in comparison to [ 89 Zr]-B12 a change was observed with reduced kidney uptake noted in deplete animals ( Fig. 3; Table S1).

Discussion
We first characterized the apo-IF's glycosylation profile from A. thaliana given the postulated differences in glycol profile for human versus plant IF, and the role such sugars can play in terms of receptor recognition and binding, and protein clearance. GC-MS data showed a plant glycosylation profile of α(1-3)-fucose, xylose, mannose and n-acetylglucosamine in the ratios 0.17: 0.18: 1.0: 0.24, respectively. Since galactose was not detected the most likely receptor causing the liver internalization of IF was the mannose receptor CD206, which recognizes fucose, mannose, and n-acetylglucosamine and is found in liver epithelial cells and macrophages 31 . ASGPR is also highly expressed in the liver, however, this receptor recognizes terminal galactose, which was not present on the hrIF used herein 32 . Since this differs from a human glycosylation profile we wanted to investigate the hypothesis that the glycosylation profile might alter IF recognition in the body (should be only recognized by CUBN) and be recognized by the CD206.
To test our hypothesis, we first completed in vitro experiments with a fluorescent B12 conjugate, B12-Cy5, synthesized previously 3 , to allow the performance of quantitative flow cytometry experiments. We confirmed our IF-B12-Cy5 functioned as endogenous IF and that it was recognized by CUBN in the CUBN+ cell line BN16. Then, we investigated uptake in J774.A1 macrophage cells (CUBN-and CD206+), which indicated that IF-B12-Cy5 recognition is IF specific and supports the GC-MS sugar profile. A prior report by Paveley et al. 30 , supports the results, with a similar shift in fluorescence observed when a fluorescent-monoclonal antibody for the CD206 receptor was incubated with J774A.1 cells. In addition, a near complete block in uptake was observed when J774A.1 cells were incubated with an excess (2 mg/mL) of mannan, which is reported to reduce CD206 mediated uptake 33 , 45 minutes prior to incubation with IF-B12-Cy5, supporting that the uptake is mediated via the CD206 receptor (Fig. 2). Also important to note that macrophages expressing high levels of CD206 have been considered as anti-inflammatory and therefore targeting this receptor could be toxic with prolonged administrations in vivo 34 . CUBN and CD206 negative cell line CHO-K1 (confirmed by Western blot-data not shown) did not show any association (Fig. S7).
Since the in vitro studies confirmed our hypothesis, we continued our investigation in vivo using PET imaging. Upon completion of the synthesis, characterization, and stability studies of [ 89 Zr]-B12 and IF-[ 89 Zr] _ B12 (Figs. S1-S3) indicated that in vivo PET imaging studies could be conducted. Initially, PET imaging was completed in nude athymic female mice on replete chow containing B12 at 1, 5, and 24 h p.i. (200-250 μCi/mouse via the tail vein) of IF-[ 89 Zr]-B12 (data not shown). As shown in Fig. 3 and Table S1, there was significant liver uptake at 5 h, which did not change over the subsequent 24 h. Overall, the highest uptake was observed in the liver, compared to the control ([ 89 Zr]-B12 alone) which showed uptake primarily in the kidneys.
To more closely examine the effects of B12 status on distribution, IF-[ 89 Zr]-B12 or [ 89 Zr]-B12 were injected into nude athymic female mice on a B12 deplete diet for 21 days and PET imaging was completed on mice 24 h p.i. For IF-[ 89 Zr]-B12 the highest uptake was seen in the liver and kidneys and did not look significantly different than mice on replete diets. However, in comparison to [ 89 Zr]-B12 a clear change was observed with higher uptake in the liver. This is in accord with the current view, that in rodents the kidneys store excess B12 24,35 . To quantify this change, biodistribution studies were conducted (Figs. 4 and 5 and Table S1).
[ 89 Zr]-B12 replete and deplete biodistribution showed significant change in uptake within the liver, kidneys, blood, pancreas, and heart (p < 0.05). The IF-[ 89 Zr]-B12 replete and deplete models showed significant change within the blood, and heart (p < 0.05). To date most B12 experiments show high uptake in the kidneys with less uptake in the liver, our data displays an altered pharmacokinetic (PK) and uptake profile for the IF-bound B12. This change in PK is most likely, in part, due to the CD206 receptor, highly expressed in the liver and macrophages 36,37 , which recognize the specific glycosylation profile of A. thaliana produced recombinant human IF.
So, while we did not set out to target the liver, this is what was ultimately observed. This observation has implications for drug delivery using plant derived IF, depending on how easily a bound B12-conjugate drug could be unloaded, and which cells took it up. Such work is on-going in our labs.
In conclusion, the absence of effect on IF uptake by endogenous B12 levels indicates that IF can allow for the use of B12 conjugate chemistry (i.e. B12 drug conjugates) while stepping out of the normal B12 'dietary' pathway dependent on TC mediated cellular uptake. This use of IF would diminish any possible risk, however unlikely (and certainly not confirmed to date), of developing B12 deficiency in subjects being dosed with B12 bioconjugates. The liver uptake seen in PET imaging and biodistribution when a radio-B12 complex of IF was administered was attributed to the terminal sugar being recognized by, most likely, the CD206 receptor, itself a major target for pharmaceutical intervention/targeting.
All animal experiments and manipulations were carried out after review by, and with the approval of, the IACUC committee at Wayne State University. All animal experiments and manipulations were carried out according to the guidelines and regulations set by the IACUC at Wayne State University, which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). IACUC Protocol # for this work was 17-07-302.