Characterization of hyaluronan , hyaluronidase PH 20 , and HA synthase HAS 2 in inflammation and cancer

Many biological activities have been reported for hyaluronan (HA, an extracellular matrix glycosaminoglycan), including regulation of inflammation. High molecular weight HA (HMW HA) is recognized to be anti-inflammatory, however, it is less clear what activity low molecular weight HA (LMW HA, the degradation catabolite of hyaluronidase) has in inflammation. Because of reagent contamination of both hyaluronidase and HA used in many studies, some reported aspects of the associated biology need reinvestigation. Many reports have shown that LMW HA is pro-inflammatory; however, some recent publications raised serious doubts and demonstrate that LMW HA and hyaluronidase PH20 are not pro-inflammatory. Endotoxin and other contaminants in the reagents used in previous reports (i.e., bovine testicular hyaluronidase [BTH] Hyal type I-S and IV-S from Sigma) may be responsible for the observed inflammation. Further investigation has shown that the most purified BTH (type VI-S from Sigma) contains no endotoxin, but has substantial level of peptidoglycan, which is also pro-inflammatory. Caution should be taken when conducting studies of inflammation using HA and hyaluronidase that may contain endotoxin and peptidoglycan, as well as other pro-inflammatory contaminants. We have characterized HA affinity to its receptor CD44. While HMW HA binds to CD44 strongly, LMW HA degraded by PH20 hyaluronidase does not bind to CD44 on the cell surface. This may explain the observation that recombinant human hyaluronidase PH20 (rHuPH20) inhibits leukocyte migration upon inflammatory stimulation by interrupting CD44 interaction with HMW HA. PEGPH20, a pegylated rHuPH20, has the same inhibitory activity on leukocyte transmigration as rHuPH20. Further investigation of leukocyte-inhibiting activity of rHuPH20 may reveal potential clinical applications, such as wound healing and arthritis. Another report which is difficult to interpret is that the ultra-high molecular weight HA produced by naked mole rat hyaluronan synthase-2 (nmrHAS2) contributes to cancer resistance, probably due to its specific anti-inflammatory activity. In our study, nmrHAS2 is cancer-promoting in human cancer cells, to a similar extent as human HAS2. The proposed cancer resistant activity of nmrHAS2 may apply selectively to its host animal species, the naked mole rat.


Introduction
The composition of hyaluronan (or hyaluronic acid, HA) consists of repeating polymeric disaccharides D-glucuronic acid and N-acetyl-D-glucosamine which are linked by a glucuronide bond [1] .HA, one of a number of major glycosaminoglycan components contained in extracellular matrix, is proposed to be involved in a great number of biological functions, including inflammation and tumor progression [2-7-] .In normal tissues, HA exists as the high molecular weight form (HMW HA, >1,000 kDa), and can be degraded by hyaluronidases into low molecular weight HA (LMW HA, 25-1,000 kDa), which can be taken up by liver and lymph node, or further degraded into smaller oligo HA (<10 kDa) and disaccharide.When dosed with hyaluronidase such as PH20 (which only works in the extracellular space), HMW HA is progressively degraded to LMW HA, but the amount degraded to oligo HA is much less in vivo.While HMW HA is purported to inhibit inflammation, it has been reported that LMW HA can stimulate inflammation [8] .However, the actual contribution of LMW HA on inflammation is under question.LMW HA was proposed to stimulate inflammation via binding to toll-like receptors TLR2 and TLR4.However, direct binding of LMW HA to TLR2 or TLR4 has never been reported, and contamination of reagents by various inflammatory factors was suspected to be responsible for the observed inflammation [9][10][11][12] .Our interest is to further characterize the activity of LMW HA, PH20 preparations, and in vivo-generated HA in stimulation of inflammation.

LMW HA and PH20 hyaluronidase are not inflammatory
It is of interest to explore the actual impact on inflammation by LMW HA, the catabolic product of PH20 hyaluronidase.In a recent publication, we have shown that LMW HA did not bind to TLR2 or TLR4 overexpressed on the cell surface, whereas LPS bound to TLR2 and TLR4 strongly and stimulated downstream NFB signaling [2] .Coincidently, LMW HA did not stimulate TLR2-or TLR4mediated NFB activity or induce cytokine MIP-2 production in murine macrophages.In vivo, LMW HA (175-350 kDa) not only failed to stimulate any cytokine/chemokine production, it also did not induce leukocyte infiltration.We thus conclude that LMW HA is not pro-inflammatory in vitro and in vivo.A highly purified hyaluronidase rHuPH20 did not show any proinflammatory property, and to our surprise, bovine testis hyaluronidases (BTH) Hyal type I-S and IV-S (Sigma) stimulated inflammation both in vitro and in vivo.Further analysis showed that BTH type I-S and IV-S contain high levels of endotoxin, a common pro-inflammatory contaminant, as well as other protein contaminants.Removal of endotoxin largely reduced the pro-inflammatory activity of BTH, indicating that endotoxin is a principal pro-inflammatory factor in the BTH preparations.These data suggest that, when conducting inflammation studies, specific attention must be given to the potential contamination of the reagents to avoid inaccurate conclusions.BTH Hyal type VI-S, the most purified hyaluronidase from Sigma, contains a very low level of endotoxin (<0.1 U/mg).However, it still induced inflammation, suggesting the existence of a biologically active contaminant other than endotoxin.Peptidoglycan is a pro-inflammatory factor which works through the TLRs [13, 14]   .Our analysis showed that BTH type VI-S contained high levels of peptidoglycan (Table 1), which activates both TLR2 and TLR4 (Figure 1).These data suggest that peptidoglycan in BTH type VI-S is the major proinflammatory factor.Since there is no reliable procedure to remove peptidoglycan from protein preparation, extreme caution and extensive experimental controls are needed when using BTH preparations to study inflammation.

rHuPH20 has anti-inflammatory potential
HA-CD44 interaction between cell-cell junctions (HA on the endothelium/stroma and the HA receptor CD44 on leukocytes) was reported to contribute to leukocyte migration [15,16] , and blocking this interaction is hypothesized to reduce or even inhibit leukocyte migration, an important step of the inflammatory response.We tested whether rHuPH20 and PEGPH20 (pegylated rHuPH20) can impact inflammation, since PEGPH20 was reported to efficiently degrade extracellular HA [17] .Indeed, rHuPH20 degraded HA on the cell surface and blocks HA-CD44 interaction (Figure 2A, B), and subsequently blocked LPSinduced leukocyte transmigration without impacting cytokine/chemokine production [2]   .
Additionally, PEGPH20 blocked LPS-induced leukocyte transmigration to the same extent as rHuPH20 (Figure 2C).These data suggest that rHuPH20 could have clinical activity in some inflammation-related disease indications, such as diabetic wound healing and idiopathic lung fibrosis.The differential pharmacokinetic properties of rHuPH20 (short half-life of 3-5 minutes in blood) and PEGPH20 (long half-life of 48-72 hours in blood) will provide multiple options to apply PH20-based HA-degrading reagents for specific disease conditions.

HA in inflammation and cancer
Hyaluronan has a major role in the function of immune cells during inflammation and inflammation-related tumorigenesis [18] .It is believed that a part of the effect of HA accumulation during tumorigenesis is to activate tumor-associated macrophages (TAMs).For example, the stromal compartments formed in HA-high tumors are preferentially enriched with TAMs, facilitated by HA-CD44 interaction [19] .PH20 degrades HA and interrupts HA-CD44 interaction, consequently inhibiting leukocyte migration during inflammation [2] , and PEGPH20 inhibits tumor growth in multiple HA-high cancer types [20,21] .It is possible that part of the anticancer activity of PEGPH20 may occur through inhibition of pro-tumorigenic immune cell trafficking into the tumor stroma by blocking HA-CD44 interaction.
Although a role of HA accumulation (resulting from upregulation of HA synthases or down regulation of hyaluronidases) in inflammation and tumorigenesis has been quite well recognized in the literature, there are contradictory reports of opposing functions of HA.HA accumulation by transcriptional up-regulation of HAS2 may reflect initiating molecular events in inflammation and cancer [22] .On the contrary, Tian et al reported that HAS2, which produces higher molecular weight HA, may possess activity which leads to inhibition of tumorigenesis in the naked mole rat [23] .To understand whether naked mole rat Has2 (nmrHAS2) and human HAS2 (hHAS2) have opposing function on tumor, we overexpressed nmrHAS2 and hHAS2 in a human pancreatic cancer AsPC1 cell and tested them in an in vivo tumor growth study.AsPC1 cells overexpressing nmrHAS2 and hHAS2 produced HA with similar molecular weight profiles (Figure 3), which differs from the reported observation that nmrHAS2 produced much higher molecular weight of HA in rat cells [23] .It is possible that nmrHAS2-produced HMW HA only exists in the skin of the naked mole rat, but not in other tissues of the animal or in the skin of other animal species.It is thus not surprising to find that nmrHAS2 overexpression failed to inhibit tumor growth in vivo, Figure 1.Peptidoglycan is pro-inflammatory.A-B.Human TLR2-or TLR4-overexpressing HEK293 cells with NFB-promoter driven SEAP reporter gene were used to test TLR-mediated NFB pathway stimulation.The SEAP reporter gene activity was graphed for TLR2 (A) and TLR4 (B).Different peptidoglycans (PGN, from Sigma) were tested.C. The mouse air pouch was challenged with peptidoglycans (PGN, 10 µg/ml) or Hyal VI-S (1 mg/ml, containing 4 µg/ml PGN) for 8 hours.Neutrophil density in the air pouch exudates was detected by FACS analysis using CD45/Ly-6G, and graphed (n=8).LTA, lipoteichoic acid.*p<0.05.which suggests that nmrHAS2 does not have anticancer activity in human cancer cells (Figure 3).On the contrary, nmrHas2 overexpression promoted AsPC1 tumor growth to the same extent as human HAS2 overexpression, indicating that nmrHAS2 can promote tumor growth similar to the results reported for overexpression of human HAS2 [22] .Based on our current observations, we believe that the anticancer activity of nmrHAS2 may be only applicable to the naked mole rat.In other animal species, including human, HAS2 gene upregulation is correlated to increase of tumor growth and cancer progression.

Reagent
In summary, our work indicates that some HA and hyaluronidase reagents may contain multiple proinflammatory contaminants (including endotoxin, peptidoglycan, or other proteins), which could lead to misinterpretation of their role in inflammation.With regards to tumorigenesis, HAS2 from both naked mole rat and human promoted tumor growth using a human cancer model, supporting the mainstream concept that HA accumulation caused by HAS2 upregulation is a protumorigenic factor in human cancer.
The endotoxin levels in the tested reagents were determined according to Endosafe ® -PTS™ Rapid Endotoxin Detection protocol (Charles River Laboratories, San Diego, CA, USA).The peptidoglycan levels were determined according to SLP Reagent kit protocol (Wako, Richmond, VA, USA).Stimulation of inflammation was determined by induction of leukocyte infiltration and/or cytokine/chemokine production in the mouse air pouch model of inflammation.

Figure 2 .
Figure 2. HA binding to CD44 is suppressed by rHuPH20.A-B.AF594-labeled HA was incubated with CD44+ MDA-MB-468 cells in the absence (A) or presence (B) of rHuPH20.The cells were then DAPI-stained and imaged.C. The mouse air pouch was challenged with LPS (1 µg/ml) in the presence or absence of rHuPH20 (2,400 U/ml) or PEGPH20 (2,400 U/ml) for 8 hours.Neutrophil density in the air pouch exudates was detected by FACS analysis using CD45/Ly-6G, and graphed (n=8).*p<0.05.

Figure 3 .
Figure 3. Naked mole rat HAS2 promotes growth of human pancreatic cancer xenograft tumors.A. Over-expression of HAS2 (nmr or human) increased HA level in AsPC1 tumors.B. Over-expression of nmrHAS2 did not change HA size in AsPC1 tumor model.C. Over-expression of nmrHAS2 promoted AsPC1 tumor growth in vivo.Pa, parental.*p<0.05.