WO1992017151A2

WO1992017151A2 - Method of administering drugs through modification of epithelial tight junctions by tumor necrosis factor

Info

Publication number: WO1992017151A2
Application number: PCT/US1992/002312
Authority: WO
Inventors: James M. Mullin
Original assignee: Lankenau Institute for Medical Research
Current assignee: Lankenau Institute for Medical Research
Priority date: 1991-03-27
Filing date: 1992-03-20
Publication date: 1992-10-15
Anticipated expiration: 1993-09-27
Also published as: AU1768692A; WO1992017151A3

Abstract

A method and system for paracellularly administering pharmacologically-active substances across mammalian epithelial barriers, comprising sequentially administering to predetermined epithelial cells of said mammalian epithelial cell sheets having receptors for tumor necrosis factor located thereon and having a homeostatic tight junctional resistance therebetween, an amount of a biologically-active substance reactive with said receptors for tumor necrosis factor in an amount effective to lower said homeostatic tight junctional resistance to permit paracellular transit of said pharmacologically-active substances across said epithelial cell sheets; and followed by exogenously administering a pharmacologically-active substance in a pharmaceutically-acceptable carrier to said predetermined mammalian epithelial barrier before said tight junctional resistance reverts to said homeostatic state.

Description

METHOD OF ADMINISTERING DRUGS THROUGH MODIFICATION OF EPITHELIAL TIGHT JUNCTIONS

BACKGROUND OF THE INVENTION

1. Field of the Invention.

This invention relates to improved methods and systems for administering pharmacologically-active agents by transepithelial (paracellular) administration techniques. More specifically, the transepithelial or paracellular administra¬ tion of pharmacologically-active substances may be accomplished through the prior administration of biologically-active substances that interact with the receptors for the cyto ine Tumor Necrosis Factor (the serum factor inducing hemorrhagic necrosis of transplantable tumors; "TNF"; also known as the macrophage hormone associated with cachexia or "cachectin") located on epithelial membranes, causing a reversible dis¬ ruption of the tight junctional strands between such epithelial cells, thereby promoting the passage of pharmacologically- active substances and solutes through the intercellular voids caused by such disruption of the tight junction between epi¬ thelial cells. The method is most significant with respect to the administration of higher molecular weight peptides, pro¬ teins, and the like, specifically the ability of peptides and proteins of interest to transit epithelial barriers and enter the blood stream without undergoing degradation (as would occur if they transited through - not between cells of the barrier) . The systems of the invention involve the sequential administra¬ tion of a biologically-active substance reactive with the TNF receptor followed by a pharmacologically active substance in a pharmaceutically acceptable carrier.

2. Statement of the Related Art.

It is known in the art that the junction between certain epithelial cells, the so-called "tight junction" or "tight junctional barrier," or "tight junctional region" may be disrupted due to various substances, including 12-0-tetradecan- oyllphorbol-13-acetate ("TPA") and other related phorbol ester compounds such as phorbol-12,13-dibutyrate (PDBU) . These effects have been noted n LLC-PK_X renal epithelial, T₈₄ and MDCK cell lines. Such disruption of the tight junction causes leakiness across the epithelial cell sheet.

Epithelial tight junction disrupting substances such as exogenously administered TPA and other phorbol esters have been shown to act when administered to either the apical (luminal) or basal-lateral (antiluminal) cellular surfaces. Transepithelial flux of D-mannitol and polyethylene glycol following the application of phorbol ester tumor-promoting agents to LLC-PKi renal epithelial cells has demonstrated that the normally occluding junctions between cells have been made leaky. Further studies have shown that higher molecular weight protein molecules such as epidermal growth factor ("EGF") (Mr=6,100), which is known not to readily cross epithelial barriers .in vivo, may be paracellularly (transepithelially) transferred after treatment of the epithelial cell sheet with a phorbol ester. Moreover, EGF in unmetabolized form had been shown to be recovered following transepithelial transit of the molecule following administration of a phorbol ester, thereby averting to a significant degree the expected occurrence of degradation products of EGF as would be expected upon transit of the molecule across untreated cell sheets. However, the action of agents such as TPA has heretofore been deemed to be irreversible, or only slowly reversible following cessation of application of another phorbol ester such as PDBU to the cell membranes.

Other compounds shown to produce enhanced trans¬ epithelial permeability include certain diacylglycerol com¬ pounds, such as 1,2-dioctanoylglycerol, 1,2-dicaprylglycerol, 1,1-dioleoylglycerol or l-oleoyl-2-acetylglycerol. Inasmuch as these agents also are activators of the protein kinaεe C signal transduction system, as are phorbol esters, it has been sug¬ gested that the changes in epithelial junctional permeability may, at least in part, be mediated through such a mechanism. Accordingly, the regulation of transepithelial permeability through protein C kinase activation has been suggested as a meanε of promoting the transit of protein hormones across the blood-brain barrier or the chordus-plexus epithelia. Unlike phorbol esters, the diacylglycerol compounds and, in particu¬ lar, 1,1-dioctanoyl glycerol, show some reversibility of the initial dissipation of transepithelial resistance. However, the diacyglycerols are endogenous compounds that would not under normal circumstances be accessible to the TNF receptors located on the cellular membranes of epithelial cells, and moreover have (as "intracellular second messengers") extremely wide ranging effects. In addition their action on cell junc¬ tions is also not rapidly reversible.

As is evident from the foregoing discussion, various biologically-active substances may elicit changes in the homeo¬ static tight junctional resistance of epithelial cell sheets to varying degrees. However, the biologically-active agents known to affect the tight junctional barrier suffer various drawbacks of irreversibility, the need for replenishment of said agent to maintain a reduced tight junctional resistance level, the incomplete reversion to the homeostatic resistance level following administration of said agent, and a lack of potency. Moreover, the phorbol esters and diacylglycerols suffer a severe drawback relating to their suspected status as car¬ cinogens and/or agents having plasma membrane disruptive activity.

SUMMARY OF THE INVENTION

An object of the present invention is to provide a method of administering pharmacologically-active substances which normally present absorption problems across epithelial membranes due to the structure, molecular weight, solubility, degradation or other physico-chemical problems associated with the absorption of such agents.

Another object of the present invention is to provide a system of administering pharmacologically-active substances which normally present absorption problems across epithelial membranes due to the structure, molecular weight, solubility, degradation or other physico-chemical problems associated with the absorption of such agents.

Yet another object of the present invention is to provide a method of sequentially lowering the tight junction resistance of epithelial cells below homeostatic levels to permit paracellular transport of exogenously administered pharmacologically-active substances.

It has been discovered that the foregoing objects are achieved in one aspect of the present invention in which a method is provided of transepithelially administering (by paracellular route) pharmacologically-active substances, preferably peptide, protein and other pharmacologically-active substances that normally present absorption problems across mammalian epithelial membranes, and most preferably, peptides and other pharmacologically-active substances of a molecular weight of about 6,000 or less, comprising:

(a) sequentially administering to predetermined epithelial cells of said mammalian epithelial membranes (or endothelial cells of endothelial membranes) , preferably human epithelial cells, having receptors for tumor necrosis factor (the serum factor inducing hemorrhagic necrosis of tranε- plantable tumors and the macrophage hormone also known as cachectin) (preferably TNF-alpha) located thereon and having a homeostatic tight junctional resistance therebetween, an amount (preferably about 20 to 100 ng TNF/2.5 x 10⁶ cells) of a biologically-active substance reactive with said receptors for tumor necrosis factor in an amount effective to lower said homeostatic tight junctional resistance to permit paracellular transit of said pharmacologically-active substances across said epithelial membranes; and

(b) followed by exogenously, preferably topically, administering a pharmacologically-active substance, in a pharmaceutically-acceptable carrier (for example water or other pharmaceutically acceptable diluents or carriers for said pharmacologically-active substance) to said predetermined mammalian epithelial cells before said tight junctional reεiε- tance reverts to said ho eostasic state, preferably within about 60 to about 150 minutes following the administration of said biologically-active substance reactive with said receptor for tumor necrosis factor.

Another aspect of the present invention discloses a method of paracellularly administering pharmacologically-active substances across mammalian epithelial membranes, comprising:

(a) sequentially administering to predetermined mammalian epithelial cells of said mammalian epithelial mem¬ branes having a homeostatic tight junctional resistance there¬ between, an amount (preferably at levels of about 10 to about 50 ng/ml or about 20 to 100 ng per 2.5 X 10⁶ cells) of tumor necrosis factor (preferably TNF-alpha) effective to lower said homeostatic tight junctional resistance to permit paracellular transit of said pharmacologically-active substances across said epithelial membranes; and

(b) followed by exogenuously administering a pharmacologically-active substance in a pharmaceutically- acceptable carrier (preferably substantially comprising water) to said predetermined mammalian epithelial cells before said tight junctional resistance reverts to said homeostasic state.

It has still further been discovered that the fore¬ going objects are achieved in one aspect of the present inven¬ tion comprising a transepithelial therapeutic system containing a pharmaceutic composition for transepithelial administration of pharmacologically-active substances to predetermined epi¬ thelial tissue, comprising:

(a) a tight junction resistance-lowering amount of tumor necrosis factor (preferably TNF-alpha) in a pharmaceu¬ tically acceptable carrier, which enhances the transepithelial permeability of component (b) ; and

(b) a pharmacologically-active substance (preferably peptides, proteins and the like and most preferably compounds of a molecular weight of leεε than 6,000 in a pharmaceutically acceptable carrier (preferably water) . BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 iε a graph depicting the dose response of TNF on transepithelial resistance.

Figure 2 is a graph depicting the delayed enhancement of transepithelial resistance by TNF.

Figure 3 is a graph depicting the effect of apical vs. basolateral TNF on transepithelial resistance.

Figure 4 iε a graph depicting the transepithelial resistance decrease induced by TNF.

Figure 5 is a graph depicting the effect of TNF on LLC-PK_! transepithelial resistance: the 4 phases of the re¬ sponse.

Figure 6 is a graph depicting the effect of TNF on LLC-PK- transepithelial resistance following pretreatment with genistein, to block tyrosine kinase.

Figure 7 is a graph of TNF on LLC-PKi transepithelial resiεtance: pretreatment with cycloheximide to block protein εyntheεiε.

Figure 8 is a graph of lack of tight junctional res- ponεe upon secondary TNF application.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

As used herein, the term "epithelial cells" iε deemed to include epithelial as well as endothelial cells lining the blood and lymph vasculature and other closed bodily cavities such as the heart and thoεe capillarieε forming the blood-brain barrier. The terms "paracellularly" and "tranεepithelially" are intended to connote transit of materials or solutes around and between epithelial cells, rather than through intracellular absorption of such materials and solutes. As signified above, the terms "tight junctional barrier," "tight junctional region" or "tight junction" are known in the art to signify the ring¬ like junctional seal between mammalian epithelial cells which line body cavities. The term "homeoεtaεiε" or "homeoεtatic" iε intended to connote the normal dynamic equilibrium εtate of the tight junctional resistance of mammalian epithelial cells. "Sequential application" of the TNF-receptor activator and the pharmacologically active substance may potentially encompass a negligible time difference to one of several hours, reflecting the different physiology of different epithelial barriers and different species, which selection may be made by one of ordinary skill in the art without undue experimentation.

An essential function of epithelial cells of the human body is to divide the body into separate fluid-filled compartments, delimited by these epithelial (or endothelial) cells. The cellε dividing these compartments perform, beyond their essential barrier function, thermodynamic work by trans¬ porting solutes unidirectionally from one fluid compartment to the other, thereby defining and maintaining the unique compo¬ sition of these compartments. The cells of the epithelial or endothelial barrier are, however, only one component of the barrier. Equally important is the junctional band surrounding each cell of the cell sheet. This "tight junction" or zonula occludenε is a semipermeable barrier which allows certain solutes (depending upon their size and charge) to pass through this paracellular pathway between the cells from one fluid compartment to the other. The tight junction is fully distinct from the zonula adherens, adhering junction, or "desmosome" which presents no barrier to solute flow. The tight junction is composed of a network of protein strands, the composition of which is just beginning to be understood. It is believed that the tight junction is not static, but appears to be highly regulated, producing permeability changes in response to a large number of changing physiological conditions. Accord¬ ingly, it is deemed to play a key role in transepithelial transport processes previously thought to occur through the cellε per εe.

Unidirectional tranεport of solutes from one fluid compartment to another frequently proceeds against a concentra¬ tion and/or electrical gradient and therefore requires energy. Factorε εuch aε εolubility, molecular weight and numerous other pharmacokinetic factors govern the tranεport of pharmacologically-active substances across cell membranes. If selected (predetermined) epithelial barriers in the body are made leaky through a decreased reεistance in tight junctional (paracellular) pathways, solutes, including pharmacologically- active substances will avert the energy intensive and pharma- cokinetic limited aspects of drug absorption and distribution.

I have discovered that the effect of the macrophage- secreted protein, tumor necrosis factor (TNF) , on epithelial tight junctions produces a marked physiological effect upon the tight junctional region between epithelial cells, thereby lowering transepithelial resistance and voltage potential difference, and increasing the flow of solute between cellε and acroεε the epithelium. These effects are dose dependent, rapidly reversible, and inhibited by a monoclonal antibody to TNF. This discovery suggests that the endogenous release of TNF at various sites throughout the body will cause a breakdown in the barrier function of the epithelial cell sheet at that locus. Moreover, my discovery demonstrates that exogenouε administration of TNF may be targeted to react with TNF re¬ ceptors on epithelial cellε of a predetermined epithelial membrane and may be therapeutically employed to modify the occlusion of epithelial cells to facilitate the transit of solutes including normally problematic pharmacologically-active substances, as aforesaid.

TNF is but one member of a family of cytokineε releaεed by cellε of the immune system in response to various invasive stimuli. In my experiments TNF-alpha has been shown to be the moεt active form of the cytokine and iε εpecified for uεe in the preferred embodiment of the invention and iε deemed to be the beεt mode of practicing the invention. TNF-beta exhibitε εignificantly leεε activity.

To demonεtrate the preferred embodiment of my method utilizing TNF, an LLC-PKi renal epithelial cell line waε choεen becauεe of itε cloεe parallel in phyεiological properties to the proximal tubule of the kidney, one of the most extenεively inveεtigated epithelial preparations. By culturing LLC-Pi cells on permeable filters, transepithelial voltage and resis¬ tance (the inverse of permeability) across the cell sheets, as well as the transepithelial flux of radiolabeled solutes, can be monitored.

With reference to Fig. 1, unlike the near immediate effects of prior art compounds such as phorbol esters, an effect of TNF on transepithelial resistance does not occur until almost 90 minutes after TNF exposure to epithelial cell membranes. The sharp drop in resistance reverses within 60 minutes and is also dose dependent. Dose responses in the 10-50-ng/ml (20 to 100 ng/2.5 X 10⁶ cells) range have been observed in in vitro testing. This same effect, and the same time course of fall and recovery of resistance, occurs whether the TNF iε removed from the culture medium 30 minuteε after application or is left in contact with the cells throughout the study.

The practical effect of the foregoing pharmacological response iε that the TNF may be administered as a single application to an epithelial membrane, whereupon once in contact with TNF, the affected cell membranes will lower the resistance of the tight junctional barrier, allowing the passage of additional solutes including, for example, exogenously administered pharmacologically-actice substances.

With reference to Fig. 2 and 8, and without suggest¬ ing that a possible mechanism of action may limit the invention disclosed and claimed herein in any manner, because the con¬ tinued presence of TNF does not block the recovery of tight junction reεiεtance and return to a homeostatic state, it iε εpeculated that the original intracellular εignal which causes transepithelial resiεtance to drop iε attenuated, or a εecond oppoεing εignal is generated. The latter mechanism is believed to play a part in the return to homeoεtaεiε becauεe when protein εyntheεiε haε been blocked prior to adminiεtration of TNF, the tight junction reεiεtance-lowering effectε of TNF are recorded, but the reversal to a homeoεtatic εtate of tenεion iε not evidenced. (See Figure 7) .

Fig. 2 exemplifieε the normal occurrence that reεiε- tance not only returnε to original valueε but also normally rises above control values. This "overshoot" is transient and resistance values return to initial levels usually before 12 hours. The observed gradual decrease in control resistance levels iε characteristic of my experiments. It is typically a slow, graded decrease which rarely falls below 80% of initial resistance values.

To confirm that the TNF-induced sharp decrease in transepithelial resistance iε due to induced leakinesε in the tight junctions, 1 mM D-["C]mannitol was added to the baso- lateral fluid compartment and then the rate of appearance of D["C]-mannitol in the apical compartment waε measured. It is underεtood in the art that D-mannitol has negligible affinity for any membrane transport system in LLC-PK_X cellε and must cross the cell sheets by passing across the tight junctions and moving between the cells.

With reference to Fig. 1, when cell sheets were treated with TNF for 2 hours at 37°C (tranεepithelial reεiε- tance had fallen to approximately 60% of initial valueε) , the unidirectional flux of 1 mM D-[¹C]mannitol acroεs the TNF- treated cell sheets had risen to 138% of control flux values, as demonstrated in Table 1 below. By 3 hours (by which time resistance had recovered to over 70% of initial values) the D-mannitol flux rate had returned to 118% of control valueε.

TABLE 1

Tranεepithelial flux of 1 mM D-[¹⁴C]mannitol across control and TNF-treated LLC-PKx cell sheets

" Average resistance of three cell sheets ± range of values. ^b Average of linear regression determinations for three cells sheets ± SE.

Unlike the phorbol ester studies known in the art, it is believed that the effects of TNF on transcellular mannitol flux would be more dramatic, except for the relatively rapid recovery of tight junction resistance, which was not true with phorbol esterε. Recovery is sufficiently rapid to be at or near completion during the mannitol flux measurements. Thus, the recorded flux valueε are actually a reflection of an average permeability of the junctionε (between many cells) when resistance fell to its lowest value, and at a later stage when resistance is recovering.

When LLC-PK-_L cells are cultured on plastic or glasε, as opposed to a permeable filter, the confluent monolayer is highlighted by the appearance of fluid-filled "domes." These domes arise from the unidirectional transport of solutes and water from the culture medium (apical compartment) to the space between cell εheet and dish (basolateral compartment) . When 50 ng/ml TNF was added to the culture medium there waε initially no effect, but by 2 hourε many domeε began to collapεe. This dome collapεe iε another manifestation of the decrease in tight junctional resiεtance cauεed by TNF, becauεe domes collapse when tight junctions allow backleak of the solutes and water transported acroεε to the underεurface. Similar effects were observed in the art when phorbol esters and diacylglycerolε were applied to cells of this type. However, TNF cauεeε a more rapid recovery of the domeε on the monolayer, with near total recovery of domes by 4 hourε. This time courεe of TNF inducing dome collapse and recovery iε very similar to the time course of the TNF-induced resistance changes deεcribed in Fig. 1. It is also noted that there iε no indication of any cell degenera¬ tion accompanying the dome collapεe, even at TNF levelε of 100 ng/ml for 24-hour incubations at 37^PC. This indicates that TNF is not cytotoxic to these epithelial cells.

Surface-mediated TNF effects, were noted when TNF waε applied to either the apical (luminal) or the baεolateral surface of LLC-PKχ cell sheets. TNF induced a decrease in resistance when presented to either the apical or the baso- lateral cell surface (See Fig. 3) . The drop in resistance was of equal magnitude and identical time course in each case. When presented simultaneously to both cell surfaces, the time course was again the same but the effect waε additive. It iε unlikely that TNF iε simply leaking acrosε the cell sheet to the surface containing its receptors because it iε a relatively large molecule. TNF monomerε have molecular weightε of 17,000 and the active form of TNF may be a trimer with a molecular weight of 55,000. Thiε signifies TNF would be active in this effect from within a luminal compartment.

To test whether TNF itself iε cauεing the obεerved changes in transepithelial resistance, TNF was preincubated with a monoclonal antibody to TNF for 30 minutes at 25°C in culture medium. This resulted in a complete inhibition of the effect of TNF on tight junctional resistance (See Fig. 4) . When incubated with cell sheets by itself, the monoclonal antibody produced no effect on transpithelial reεiεtance.

Sporadic or prolonged diεεipation of tranεepithelial gradientε and tranεepithelial εolute leakε in different regions of epithelial cellε can be avoided becauεe: (1) εub-optimal levelε of TNF would be without effect and (2) repeated expoεure to TNF iε without εecondary effect (See Figure 8) . A level of TNF of about 20 to 100 ng per 2.5 X 10⁶ cellε haε proven εatiε- factory to elicit the deεired tight junction modifying effect.

Four phaεeε to the TNF (tumor necroεiε factor) - induced increase in permeability of the LLC-PK, epithelial barrier (i.e. the decrease in transepithelial resistance) are now known (see Figure 5) : a delay phase (60 - 90 mins) , the drop in resistance (approx 30 mins) , the recovery of resistance (30 - 180 mins) , and the increase of resistance above initial levels.

These phases are not only physiologically distinct (and reproducible with similar time course) but show distinct biochemical regulation. The tyrosine kinase inhibitor, genis- tein, is to date the only compound which I have tested which can block (dose dependently) the actual effect, i.e. the increase in junction permeability (the decrease in trans¬ epithelial resistance) (see Figure 6) . This indicates that events in the delay and/or decrease phase are governed at least in part by a tyrosine kinase.

My initial hypothesis concerning the 60 - 90 min delay for TNF to cause a junction opening was that proteins necessary for this effect are being synthesized immediately after TNF contacts its receptors. However, protein synthesis was shown not to be necessary for the junctions to open, but it is neceεsary for the junctions to close (see Figure 7) . Cycloheximide, a classic protein syntheεiε inhibitor, does not block the resistance decrease but does block the resistance recovery.

I have demonstrated that if one applies TNF to a epithelial barrier shortly (2 - 6 hrs) after resistance iε recovering, a second opening of junctions doeε not enεue (εee Figure 8) . This means that after the initial TNF effect occurs, there is a period of refractility whereby further exposure to TNF will not elicit a second round of response.

The clinical, pharmaceutic significance to theεe observationε are aε follows:

1. An epithelial barrier will not stay chronically "open" if TNF iε in continual contact, or if TNF iε readminis- tered after a fixed period of time. This means that the physiological side effects should be minimized as e.g. residual administered TNF cannot keep the barrier leaky.

2. The TNF action on epithelial barriers appears to be capable of being fully blocked by inhibiting tyrosine kinase. Thiε may allow εelectivity in tissue specific action.

3. The TNF effect on epithelial barriers will not be reversible if the ability of the cells to synthesize new protein is compromised. This signifies that use of the TNF receptor aε a means of drug delivery must preclude the simul¬ taneous use of any drugε which would impair protein synthesis.

The invention will be further clarified by consid¬ eration of the following examples, which are intended to be purely exemplary of the use of the invention.

EXAMPLES

The pig kidney epithelial (LLC-PK,) cells used in these studies are from passages 185 to 200. Routine culturing entailed seeding 1 X 10⁵ cells in a 75-cm^z culture dish con¬ taining 25 ml of alpha-minimum essential medium with 10% fetal bovine serum. After 1 week at 37°C and 5% C0₂, the culture reached confluence, it waε then trypsinized, and the passaging was repeated.

For measurementε of transepithelial resiεtance and potential difference (voltage) , a trypsinized suspension was seeded into Millicell HA future cup assemblies (Millipore Corp.) containing 2 ml of culture medium with 10% fetal bovine serum and incubated at 37°C in a 5% C0₂ humidified atmosphere. Theεe 30mm-diameter aεεemblies had a filter base with 0.45-um poreε. Three such filter-cup assemblieε were placed in a 100-mm Petri diεh containing 15 ml of culture medium. When cellε formed a confluent monolayer, an intact epithelium exists across the filter base. Medium in the ring becomes the apical (luminal) fluid compartment and medium in the Petri dish becomeε the baεolateral fluid compartment. Three days after seeding, the spontaneous transepithelial voltage iε measured with 3 M KCl/agar bridges in series with calomel electrodes connected to a Fluk 802OB multimeter. Resistance measurements were performed by passing 40-uA current pulses via silver/AgCl electrodes connected to the culture medium via a second pair of 3 M KCl/agar salt bridges. Voltage deflections were measured with a third set of salt bridges connected to calomel elec¬ trodes, which in turn were connected to a Keithley model 197 autoranging digital multimeter.

For measuring transepithelial fluxes of D-( C) mannitol, cell sheets in filter-cup assemblies were refed with normal medium or medium containing TNF. Resistance values were measured prior to the above refeeding and then at 0.5-hour intervals. After 2 hourε at 37"C, resistance values were maximally decreased. The control set (of three cell sheets) and one set of TNF-treated cell sheets were then rinsed three timeε in (morpholinopropanesulfonic acid-buffered) saline at 25^βC. Sets of three cell sheets were incubated in Petri dishes at 25°C with 2 ml of apical saline, and 15 ml of basolateral saline containing 1 mM D(¹⁴C) mannitol. Samples of apical saline (50 ul) were removed from each filter cup at 30-min intervals, and the radioactivity was determined by liquid scintillation counting. The radioactivity (dpm) was converted to umol of D-mannitol, by dividing by the specific activity. The appearance of radiolabeled mannitol was then plotted aε a function of time. The rate of D-mannitol flux waε determined by linear regression.

The culture medium used in these εtudieε is a product of Hazelton Research Products. The fetal bovine serum was purchased from Hy-Clone Laboratories, Inc. The human recom- binant tumor necrosis factor-alpha waε obtained either from Boehringer Mannheim [2.2 X 10⁷ unitε/mg protein (determined by a cell lytic aεsay with L929 cultures)] or Genentech [4.3 X 10⁷ units/mg protein] . The anti-TNF-alpha monoclonal antibody waε a product of Boehringer Mannheim. The D-[¹⁴C] mannitol uεed in the tracer flux study was obtained from ICN Radiochemicalε, Inc. Example 1

Cell εheetε in filter-cup assemblies were refed with normal medium or medium containing 40 ng/ml TNF. Resistance values were measured at 0.5-hour intervalε, until maximally de¬ creased (2 hourε) , and then cell sheets (Sets A and B) were rinsed in saline and a flux experiment was performed as de¬ scribed above. The rate of D-mannitol flux was then determined by linear regression of the plots of appearance of D-["C] mannitol versus time. The third set of cell sheets (Set C) waε incubated for 3 hours at 37^βC, by which time resistance re¬ covered to 72% of the initial readings. The above flux proce¬ dure waε then performed on these cell sheets. The results are reported above in Table 1.

Example 2

The effect of various concentrations of TNF on trans¬ epithelial resistance across LLC-P i cell sheetε waε εtudied. Resistance measurements were taken at time zero acrosε cell sheets refed with fresh control medium 4 hours previously. Five εetε of cell εheetε (in filter cupε) were then refed with control medium or medium containing variouε concentrationε of human recombinant TNF-alpha. Reεiεtance meaεurementε were taken at the 0.5-hour intervals shown, for a period of 7 hours. Results are expressed aε the percentage of the time zero reading (100%) of each cell sheet. One representative cell sheet (of a set of three) iε shown for each condition. Spon¬ taneous transepithelial voltage readings were taken before each resistance reading and similar traces to the reεiεtance data were obtained. Reεultε of the reεiεtance meaεurementε are presented graphically in Fig. 1.

Example 3

The transepithelial resistance across pairs of LLC-PK_: cell εheetε waε meaεured following the general proce- dureε outlined above. At time zero one set of cell sheetε (control) in filter-cup aεεemblies waε refed with normal medium. The other εet waε refed with medium containing 10 ng/ml TNF. The readings of two cell sheets for each condition are reported here. The control resistances show a slight (15%) but steady decline of resistance values during the course of the experiment. Cell εheetε treated with TNF have the charac¬ teristic reversible resistance drop at approximately 2 hours. After the recovery, resistance values not only rise above control levels but eventually exceed initial levelε before slowly returning to control levels. Results of this experiment are presented graphically in Fig. 2.

Example 4

The effect of apical versus basolateral expoεure of LLC-PKj cell sheets to TNF. After initial resistance readings as described in Fig. 1, cell sheets were refed with normal medium in both apical and basolateral compartments, medium containing 10 ng/ml TNF in both compartments, or normal medium in one compartment and TNF-containing medium in the other compartment. Resistances were recorded at the times indicated aε described in "Materials and Methods." The readings of two cell sheets for each condition are shown here. The charac¬ teristic resistance drop at 2 hours occurred whether the TNF was in the apical or basolateral compartment. When placed in both compartment , the effect was additive. Experimental results are presented graphically in Fig. 3.

Example 5

This experiment demonstrates that the anti-TNF-alpha monoclonal antibody inhibits TNF-induced decrease in tranε¬ epithelial resistance. Preincubating 10 ng/ml TNF-alpha (330 unitε/ml) with anti-TNF-alpha (150 unitε/ml) in culture medium for 30 min at room temperature completely blocked the reεiε¬ tance decreaεe cauεed by TNF-alpha. At time zero, culture medium (control) , culture medium with TNF-alpha, or culture medium with THF-alpha preincubated with anti-TNF-alpha antibody waε added to the apical compartment of sets of three cell sheetε. Control medium waε added to the baεolateral compart¬ ment. Reεiεtance waε meaεured at the timeε indicated for a period of 7 hourε. Reεultε are expreεεed aε the average of three cell sheetε (percertage of initial resistance) ±SE. Cell sheetε incubated with antibody but without TNF-alpha showed a tracing very similar to that of control.

Other embodiments of the invention will be apparent to those skilled in the art from a consideration of this specification or practice of the invention disclosed herein. It is intended that the specification and examples be con¬ sidered aε exemplary only, with the true scope and spirit of the invention being indicated in the following claims.

Claims

What iε claimed iε:

1. A method of paracellularly administering pharmacologically-active substances across mammalian epithelial barriers, comprising:

(a) sequentially administering to predetermined epithelial cells of said mammalian epithelial cell sheets having receptors for tumor necrosis factor located thereon and having a homeostatic tight junctional resistance therebetween, an amount of a biologically-active substance reactive with said receptors for tumor necrosis factor in an amount effective to lower said homeoεtatic tight junctional resistance to permit paracellular transit of said pharmacologically-active sub¬ stances acrosε said epithelial cell sheets; and

(b) followed by exogenously administering a pharmacologically-active substance in a pharmaceutically- acceptable carrier to said predetermined mammalian epithelial barrier before said tight junctional resistance reverts to said homeostasic εtate.

2. The method of claim 1, wherein said biologically-active substance is tumor necrosiε factor (alpha form) .

3. The method of claim 1, wherein said mammalian epithelial cellε are human epithelial cellε.

4. The method of claim 1, wherein said exogenous administration of said pharmacologically-active subεtance iε by topical adminiεtration to εaid predetermined mammalian cellε.

5. The method of claim 1, wherein said adminiεtra¬ tion iε by oral adminiεtration.

6. The method of claim 1, wherein εaid pharmaceuti¬ cally-acceptable carrier compriεeε water.

7. The method of claim 1 wherein said sequential administration of said exogenously administered pharmacologically-active subεtance occurε from about 60 to about 150 minutes following εtep (a) of εaid proceεε.

8. A method of paracellularly administering pharmacologically-active substanceε across mammalian epithelial membranes, comprising:

(a) sequentially administering to predetermined mammalian epithelial cells of said mammalian epithelial bar¬ riers having a homeostatic tight junctional resistance there¬ between, an amount of tumor necrosiε factor effective to lower said homeostatic tight junctional resistance to permit para¬ cellular transit of said pharmacologically-active substances across εaid epithelial membranes; and

(b) followed by exogenuously administering a pharmacologically-active substance in a pharmaceutically- acceptable carrier to said predetermined mammalian epithelial cells before said tight junctional resistance reverts to said homeostaεic εtate.

9. The method of claim 8, wherein εaid mammalian epithelial cells are human epithelial cellε.

10. The method of claim 8, wherein εaid exogenous administration of said pharmacologically-active substance iε by topical administration to said predetermined mammalian cells.

11. The method of claim 8, wherein said topical administration iε by oral administration.

12. The method of claim 8, wherein εaid pharmaceuti¬ cally-acceptable carrier is water.

13. The method of claim 8, wherein εaid sequential adminiεtration of εaid exogenouεly adminiεtered pharmacologically-active εubstance occurs about 60 to about 120 minutes following step (a) of said process.

14. A tranεepithelial therapeutic εyεte containing a pharmaceutical compoεition for tranεepithelial adminiεtration of pharmacologically-active εubεtanceε to predetermined epi¬ thelial tissue, comprising:

(a) a tight junction reεiεtance-lowering amount of tumor necroεiε factor in a first pharmaceutically-acceptable carrier, which enhances the transepithelial permeability of component (b) ; and

(b) a pharmacologically-active substance in a second pharmaceutically-acceptable carrier.

15. The system of claim 14, wherein said pharmacologically-active substance is a protein or peptide.

16. The system of claim 14, wherein said first pharmaceutically-acceptable carrier comprises water.

17. The system of claim 14, wherein said predeter¬ mined epithelial tissue is selected from the group consisting of rectal, intranasal, alimentary, endothetral and corneal epithelial tissue.

18. The system of claim 14 wherein said pharmaco¬ logically active substance is water soluble.

19. The system of claim 14, wherein said second pharmaceutically-acceptable carrier comprises water.

20. The system of claim 1, wherein said predeter¬ mined epithelial tissue-is selected from the group consisting of rectal, intranasal, alimentary, endothelial and corneal epithelial tissue.