[CANCER RESEARCH (SUPPL.) 50, 814s-819s, February I, 1990)
`
`This material may be protected by Copyright law (Title 17 U.S. Code)
`
`Physiological Barriers to Delivery of Monoclonal Antibodies and Other
`Macromolecules in Tumors'
`Rakesh K. Jain
`Department of Chemical Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213-3890
`
`the microvascular wall; and (c)
`across
`transport
`space;
`(b)
`transport
`through the interstitial space. Each of these transport
`processes may involve convection (i.e., solute movement asso-
`ciated with bulk solvent movement) and diffusion (i.e., solute
`movement
`resulting from solute concentration
`gradients).
`In
`addition, during this journey the molecule may bind nonspecif-
`icaIly to proteins or other tissue components, bind specifically
`to the target(s), and/or be metabolized (for review, see Refs. 3
`and 4). In this article, I will critically review these physiological
`barriers
`to macromolecular
`transport
`in tumors
`and discuss
`some strategies to overcome them for therapeutic benefit.
`
`Distribution through Vascular Space
`
`Abstract
`
`The efficacy in cancer treatment of monoclonal antibodies or other
`macromolecules bound to radionuclides, chemotherapeutic agents, toxins,
`enzymes, growth factors, or effector antibodies has been limited by their
`inability to reach their target in vivo in adequate quantities. Heterogeneity
`of tumor-associated antigen expression alone has failed to explain the
`nonuniform uptake of antibodies. As a result, only in recent years have
`the peculiarities of tumor physiology been recognized as determinants of
`antibody distribution. Three physiological barriers responsible for the
`poor localization of macromolecules in tumors have been identified: (a)
`heterogeneous blood supply;
`(b) elevated interstitial pressure; and (c)
`large transport distances in the interstitium. The first barrier limits the
`delivery of blood-borne molecules to well-perfused regions of a tumor;
`the second barrier reduces extravasation of fluid and macromolecules in
`the high interstitial pressure regions and also leads to an experimentally
`verifiable, radially outward ccnvecrion in the tumor periphery which
`opposes the inward diffusion; and the third barrier increases the time
`required for slowly moving macromolecules to reach distal regions of a
`tumor. Binding of antibody to an antigen further
`lowers the effective
`diffusion rate of the antibody by reducing the amount of mobile antibody.
`Due to micro- and macroscopic heterogeneities
`in tumors,
`the relative
`magnitude of each of these barriers would vary from one location to
`another and from one day to the next in the same tumor and from one
`If the genetically engineered macromolecules, e.g.,
`tumor to another.
`lymphokines, and other new modalities, e.g., killer lymphocytes, as well
`as low molecular weight cytotoxic agents, arc to fulfill
`their clinical
`promise, methods must be developed to overcome these physiological
`barriers. Some of these methods are discussed, and situations wherein
`these barriers may not be a problem arc pointed out.
`
`The tumor vasculature consists of (a) vessels recruited from
`the preexisting network of the host vasculature and (b) vessels
`resulting from the angiogenic response of host vessels to cancer
`cells (5, 6), Movement of molecules through the vasculature is
`governed by the vascular morphology (i.e., the number,
`length,
`diameter, and geometrical arrangement
`of various blood ves-
`sels) and the blood flow rate.
`Vascular Morphology. Although the tumor vasculature orig-
`inates from the host vasculature,
`its organization may be com-
`pletely different depending upon the tumor
`type,
`its growth
`rate, and its location. The architecture
`is different not only
`among various tumor
`types but also between a spontaneous
`tumor and its transplants
`(for review, see Ref. 7).
`Macroscopically,
`the tumor vasculature can be studied in
`terms of two idealized categories: peripheral
`and central.
`In
`tumors with peripheral vascularization,
`the centers are usually
`poorly perfused (Fig. I). In those with central vascularization,
`one would expect the opposite. Hence,
`the penetration of blood
`borne substances should follow the same pattern.
`In reality, a
`tumor may consist of many territories, each exhibiting one or
`the other of these two types of idealized vascular patterns.
`Microscopically,
`the tumor vasculature is highly heteroge-
`neous and does not conform to the standard normal vascular
`organization (i.e., artery to arteriole to capillaries to postcapil-
`lary venule to venule to vein). Based on their ultrastructure,
`the
`(a) arteries
`tumor vessels can be classified into nine categories:
`and arterioles;
`(b) nonfenestrated
`capillaries;
`(c) fenestrated
`(d) discontinuous
`capillaries (sinusoids);
`(e) blood
`capillaries;
`lining; (f) capillary sprouts;
`channels without endothelial
`(g)
`postcapillary venules (giant capillaries);
`(h) venules and veins;
`and (i) arteriovenous anastomoses (shunts) (7). Note that except
`for vessels of classes e and f, the remaining vessel
`types are
`structurally
`similar
`to those found in a normal
`tissue. The
`vessels of classes e and f are found in healing (granulation)
`tissue. A key difference between normal and tumor vessels is
`that
`the latter
`are dilated,
`saccular, and tortuous
`and may
`contain tumor cells within the endothelial
`lining of the vessel
`wall (7). In addition, unlike a normal
`tissue with a fixed route
`between arterial
`and venous sides, a tumor may have blood
`flowing from one venule to another via vessels of classes b
`through g, or directly via an arteriovenous
`shunt. Furthermore,
`due to the peculiar nature of the vasculature,
`the organization
`814s
`
`Introduction
`The advent of hybridoma technology and genetic engineering
`has led to the design and large-scale production of monoclonal
`antibodies and other biological macromolecules potentially use-
`ful for cancer detection and treatment. These molecules can be
`conjugated to radionuclides, chemotherapeutic
`agents,
`toxins,
`growth factors,
`enzymes, effector antibodies, or
`liposomes.
`Moreover, a number of macromolecular
`cytolytic molecules
`(e.g., tumor necrosis factor,
`Iymphokines) and killer cells are
`under active investigation as potential
`therapeutic agents. While
`the concept of using antibodies with a high degree of specificity
`for tumor-associated
`antigens
`remains attractive
`for cancer
`therapy, clinical results have not, to date, lived up to the earlier
`promises of their perceived potential
`(for review, see Refs. 1
`and 2), A key problem with antibodies and other macromole-
`cules is their inability to reach all regions of a tumor in adequate
`quantities. Heterogeneity
`of tumor associated antigen alone
`cannot account for the poor penetration of antibodies in tumors.
`What mechanisms then are responsible for this maldistribution
`of monoclonal antibodies and other macromolecules
`in tumors?
`A blood-borne molecule that enters the tumor circulation
`reaches cancer cells via: (a) distribution through the vascular
`
`and Ra-
`the "Second Conference on Radioimmunodetection
`I Presented at
`of Cancer, .. September 8-10, 1988, Princeton, NJ. This article
`dioimmunotherapy
`is based on research supported by grants from the National Cancer Institute, The
`National Science Foundation,
`the American Cancer Society, and the R. K. Mellon
`Foundation; by an NIH Research Cancer Development Award; and by a Guggen-
`heim Fellowship.
`
`1
`
`

`

`[CANCER RESEARCH (SUPPL.) 50, 814s-819s, February I, 1990)
`
`Physiological Barriers to Delivery of Monoclonal Antibodies and Other
`Macromolecules in Tumors'
`
`Rakesh K. Jain
`Department of Chemical Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213-3890
`
`the microvascular wall; and (c)
`across
`transport
`space;
`(b)
`transport
`through the interstitial space. Each of these transport
`processes may involve convection (i.e., solute movement asso-
`ciated with bulk solvent movement) and diffusion (i.e., solute
`movement
`resulting from solute concentration
`gradients).
`In
`addition, during this journey the molecule may bind nonspecif-
`icaIly to proteins or other tissue components, bind specifically
`to the target(s), and/or be metabolized (for review, see Refs. 3
`and 4). In this article, I will critically review these physiological
`barriers
`to macromolecular
`transport
`in tumors
`and discuss
`some strategies to overcome them for therapeutic benefit.
`
`Distribution through Vascular Space
`
`Abstract
`
`The efficacy in cancer treatment of monoclonal antibodies or other
`macromolecules bound to radionuclides, chemotherapeutic agents, toxins,
`enzymes, growth factors, or effector antibodies has been limited by their
`inability to reach their target in vivo in adequate quantities. Heterogeneity
`of tumor-associated antigen expression alone has failed to explain the
`nonuniform uptake of antibodies. As a result, only in recent years have
`the peculiarities of tumor physiology been recognized as determinants of
`antibody distribution. Three physiological barriers responsible for the
`poor localization of macromolecules in tumors have been identified: (a)
`heterogeneous blood supply;
`(b) elevated interstitial pressure; and (c)
`large transport distances in the interstitium. The first barrier limits the
`delivery of blood-borne molecules to well-perfused regions of a tumor;
`the second barrier reduces extravasation of fluid and macromolecules in
`the high interstitial pressure regions and also leads to an experimentally
`verifiable, radially outward ccnvecrion in the tumor periphery which
`opposes the inward diffusion; and the third barrier increases the time
`required for slowly moving macromolecules to reach distal regions of a
`tumor. Binding of antibody to an antigen further
`lowers the effective
`diffusion rate of the antibody by reducing the amount of mobile antibody.
`Due to micro- and macroscopic heterogeneities
`in tumors,
`the relative
`magnitude of each of these barriers would vary from one location to
`another and from one day to the next in the same tumor and from one
`If the genetically engineered macromolecules, e.g.,
`tumor to another.
`lymphokines, and other new modalities, e.g., killer lymphocytes, as well
`as low molecular weight cytotoxic agents, arc to fulfill
`their clinical
`promise, methods must be developed to overcome these physiological
`barriers. Some of these methods are discussed, and situations wherein
`these barriers may not be a problem arc pointed out.
`
`The tumor vasculature consists of (a) vessels recruited from
`the preexisting network of the host vasculature and (b) vessels
`resulting from the angiogenic response of host vessels to cancer
`cells (5, 6), Movement of molecules through the vasculature is
`length,
`governed by the vascular morphology (i.e., the number,
`diameter, and geometrical arrangement
`of various blood ves-
`sels) and the blood flow rate.
`Vascular Morphology. Although the tumor vasculature orig-
`inates from the host vasculature,
`its organization may be com-
`pletely different depending upon the tumor
`type,
`its growth
`rate, and its location. The architecture
`is different not only
`among various tumor
`types but also between a spontaneous
`tumor and its transplants
`(for review, see Ref. 7).
`Macroscopically,
`the tumor vasculature can be studied in
`terms of two idealized categories: peripheral
`and central.
`In
`tumors with peripheral vascularization,
`the centers are usually
`poorly perfused (Fig. I). In those with central vascularization,
`one would expect the opposite. Hence, the penetration of blood
`borne substances should follow the same pattern.
`In reality, a
`tumor may consist of many territories, each exhibiting one or
`the other of these two types of idealized vascular patterns.
`Microscopically,
`the tumor vasculature is highly heteroge-
`neous and does not conform to the standard normal vascular
`organization (i.e., artery to arteriole to capillaries to postcapil-
`lary venule to venule to vein). Based on their ultrastructure,
`the
`tumor vessels can be classified into nine categories:
`(a) arteries
`and arterioles;
`capillaries;
`(b) nonfenestrated
`(c) fenestrated
`capillaries;
`(d) discontinuous
`capillaries (sinusoids);
`(e) blood
`lining; (f) capillary sprouts;
`channels without endothelial
`(g)
`postcapillary venules (giant capillaries);
`(h) venules and veins;
`and (i) arteriovenous anastomoses (shunts) (7). Note that except
`for vessels of classes e and f, the remaining vessel
`types are
`structurally
`similar
`to those found in a normal
`tissue. The
`vessels of classes e and f are found in healing (granulation)
`tissue. A key difference between normal and tumor vessels is
`that
`the latter
`are dilated,
`saccular, and tortuous
`and may
`contain tumor cells within the endothelial
`lining of the vessel
`wall (7). In addition, unlike a normal
`tissue with a fixed route
`between arterial
`and venous sides, a tumor may have blood
`flowing from one venule to another via vessels of classes b
`shunt. Furthermore,
`through g, or directly via an arteriovenous
`due to the peculiar nature of the vasculature,
`the organization
`814s
`
`Introduction
`The advent of hybridoma technology and genetic engineering
`has led to the design and large-scale production of monoclonal
`antibodies and other biological macromolecules potentially use-
`ful for cancer detection and treatment. These molecules can be
`conjugated to radionuclides, chemotherapeutic
`agents,
`toxins,
`growth factors,
`enzymes, effector antibodies, or
`liposomes.
`Moreover, a number of macromolecular
`cytolytic molecules
`Iymphokines) and killer cells are
`(e.g., tumor necrosis factor,
`under active investigation as potential
`therapeutic agents. While
`the concept of using antibodies with a high degree of specificity
`for tumor-associated
`antigens
`remains attractive
`for cancer
`therapy, clinical results have not, to date, lived up to the earlier
`promises of their perceived potential
`(for review, see Refs. 1
`and 2), A key problem with antibodies and other macromole-
`cules is their inability to reach all regions of a tumor in adequate
`quantities. Heterogeneity
`of tumor associated antigen alone
`cannot account for the poor penetration of antibodies in tumors.
`What mechanisms then are responsible for this maldistribution
`of monoclonal antibodies and other macromolecules
`in tumors?
`A blood-borne molecule that enters the tumor circulation
`reaches cancer cells via: (a) distribution through the vascular
`
`and Ra-
`the "Second Conference on Radioimmunodetection
`I Presented at
`of Cancer, .. September 8-10, 1988, Princeton, NJ. This article
`dioimmunotherapy
`is based on research supported by grants from the National Cancer Institute, The
`National Science Foundation,
`the American Cancer Society, and the R. K. Mellon
`Foundation; by an NIH Research Cancer Development Award; and by a Guggen-
`heim Fellowship.
`
`2
`
`

`

`Convection
`
`Necrotic
`Region
`
`Semi-
`Necrotic
`Region
`
`Well Vascularized
`Region
`
`Fig. l. Schematic of a tumor showing well vascularized periphery; semine-
`erotic, intermediate zone; and an avascular, necrotic central region. Low intersti-
`tial pressure in the periphery permits adequate extravasation of fluid and macro-
`molecules. These macromolecules move towards the center by the slow process
`of diffusion, which is further
`retarded by extravascular binding.
`In addition,
`interstitial fluid oozing out from tumor carries macromolecules with "it by con-
`vection into the normal tissue.
`
`TUMOR PHYSIOLOGY AND ANTIBODY DELIVERY
`tracer from a single or a limited number of regions of the tumor.
`Due to noticeable spatial and temporal heterogeneity in tumor
`blood supply,
`these values may not be representative
`of the
`whole tumor. A limited number of studies,
`in which the blood
`flow rate of the whole tumor has been measured,
`shows that
`the average perfusion rate of carcinomas
`is less than that of the
`host
`tissue or the tissue of origin. Sarcomas and lymphomas
`have higher perfusion rates than carcinomas
`(for a review, see
`Ref. 10). In general, as tumors grow larger,
`they may develop
`necrotic foci, and as a result,
`the average perfusion rate de-
`creases with tumor
`size (10). Note that even in these large
`necrotic tumors, antibody would be delivered in the well per-
`fused regions.
`Since the seminal work of Ide et al. (11), several investigators
`have examined the microscopic flow heterogeneities of tumors
`grown in transparent windows. Blood flow in tumor vessels has
`been found to be intermittent. There are random periods of
`flow reduction
`and stasis
`followed by resumption
`of flow,
`sometimes in the opposite direction (12, 13). These fluctuations
`may result
`from (a) vasomotor activity of the host arterioles;
`(b) respiratory or cardiac cycle; (e) passage of RBe, WBe, or
`cancer cells in a vessel; (d) low perfusion pressures in tumor
`vessels, and/or
`(e) elevated interstitial pressure in tumors (for
`review, see Ref. 7).
`Quantitative
`studies on the macroscopic spatial heterogenei-
`ties in the tumor perfusion rate as a function of tumor growth
`(size) are limited. Based on perfusion rates four regions can be
`recognized in a tumor:
`(a) an avascular, necrotic region; (b) a
`seminecrotic
`region;
`(e) a stabilized microcirculation
`region;
`and (d) an advancing front. In a rhabdomyosarcoma
`grown in
`the transparent
`chamber
`in a rat,
`the widths of the stabilized
`region and the advancing front were found to remain constant,
`while the widths of the necrotic and the seminecrotic zones
`increased with tumor growth. In addition,
`the perfusion rate in
`the tumor periphery ii.e., the stabilized and advancing zones)
`was found to be higher
`than that
`in the surrounding normal
`tissue (12). Intratumor
`blood flow distributions
`in spontaneous
`animal and human tumors are now being investigated using
`nuclear magnetic resonance and positron emission tomography.
`While limited,
`these results are in concert with the transplanted
`tumor studies: blood flow rates .in necrotic/seminecrotic
`regions
`of tumors
`are low, while those in nonnecrotic
`regions are
`variable and substantially higher
`than in surrounding/contra-
`lateral host normal
`tissues (14,15). As a result of these spatial
`and temporal
`heterogeneities
`in blood supply coupled with
`variations in the vascular morphology at both macroscopic and
`microscopic levels, it is not surprising that
`the spatial distribu-
`tion of macromolecules
`in tumors
`is heterogeneous
`and the
`average uptake decreases with an increase in tumor weight.
`
`from one location to another and
`of vessels may be different
`from one time to the next. As a result, one would expect
`different routes for blood flow in the well perfused advancing
`zone, seminecrotic zone, and necrotic zone (Fig. 1).
`Following the pioneering studies of Algire and Chalkley (8),
`several investigators have measured morphometric
`parameters
`of vessels in thin, two-dimensional
`tumors grown in transparent
`windows. The pioneering work of Gullino and Grantham (9)
`led to similar
`studies in three-dimensional
`experimental
`and
`human tumors
`(for review, see Ref. 7). The vascular
`space in
`tumors varies from 1 to 20% depending upon the tumor
`type,
`weight, and method of measurement.
`Studies
`in two-dimen-
`sional tumors show that vascular volume,
`length, and surface
`area increase during the early stages of growth and then de-
`crease; this behavior correlates with the onset of necrosis. The
`frequency of large diameter vessels increases in the later stages
`of growth. Most quantitative
`studies in three-dimensional
`tu-
`in vascular
`mors miss the early growth period of increase
`volume, length, and surface area. While studies of later stages
`of growth show an increase in the intercapillary distance and a
`decrease in vessel length and surface area, the results on vascular
`volume are inconclusive. Some studies show that
`the fractional
`vascular volume of
`tumors
`remains
`fairly constant
`during
`growth (suggesting an increase in the number of blood vessels
`with sluggish flow), while others show that
`the fractional vas-
`cular volume decreases as a tumor grows (in agreement with
`Once a blood-borne molecule has reached an exchange vessel,
`the observation that tumor perfusion rate decreases as a tumor
`Is (g/s), occurs by diffusion and convection
`its extravasation,
`grows) (for a review, see Ref. 7). Possible
`reasons
`for this
`and, to some extent, by transcytosis. Diffusion is proportional
`discrepancy include errors associated with different measure-
`to the surface area, S (ern"), of the exchange vessel and the
`ment
`techniques as well as presence of arteriovenous
`shunts
`and blood vessels with stagnant blood in them. Whether
`difference between the plasma and interstitial
`concentrations
`the
`to the rate of fluid
`(Cp - Ci; g/ml). Convection is proportional
`vascular volume decreases br not, a reduction in vascular surface
`leakage, IF (ml/s),
`from the vessel. IF, in turn,
`is proportional
`area would lead to a reduction in the transvascular
`exchange of
`to S and the difference between the vascular and interstitial
`molecules. In addition, an increase in the intercapillary distance
`t». - Pi, mm Hg) minus the difference
`hydrostatis pressures
`would require the molecules to traverse longer distances in the
`interstitium to reach an regions of a tumor.
`between the vascular and interstitial osmotic pressures (1l"v - 1f'i,
`mm Hg). The proportionality
`constant which relates translu-
`Blood Flow Rate. Most
`investigators
`have measured local
`(C, - e
`blood flow rate of tumors based on uptake or clearance of a minal diffusive flux to concentration
`gradients
`is
`815s
`
`Transport across Microvascular Wall
`
`j)
`
`3
`
`

`

`TUMOR PHYSIOLOGY AND ANTIBODY DELIVERY
`'40 r-----------------~
`Normal Tissue
`120
`100
`80
`60
`40
`20
`o
`
`n=1t
`
`n=7
`
`n=IO
`
`n=6
`
`n :6
`
`o-a
`
`n:5
`
`o-e
`
`Large
`
`Tumor
`
`~IO
`I
`
`EE
`
`Pxl08
`em/sec
`
`PlII08
`cmlsee
`
`Tumor Tissue
`
`140
`120
`100
`
`80
`SO
`40
`20
`o
`
`n:IO
`
`n:9
`
`n:9
`
`n=11
`
`Control Saline
`
`LGlucoseJ
`0.2
`2.0
`
`g/kO
`
`n=8
`
`n"ll
`
`lGOlaetoseJ
`02
`2.0
`
`Q/kQ
`
`n:8
`
`n"S
`
`lHyperthermiaJ
`4"
`50'
`
`J 50,000
`Fig. 2. Effective microvascular permeability coefficient (P) of M,
`dextran in normal (mature granulation) and neoplastic (VX2 carcinoma) tissues
`under various conditions: control; following saline (I mlfkg body weight) injec-
`tion; following glucose injections; following galactose injections; and following
`hyperthermia for I h (37) (mean ± SD; n = number of measurements). (From
`Ref. 8; reproduced with permission.)
`
`referred to as the vascular permeability, P (cm/s), and the
`constant which relates fluid leakage to pressure gradients
`is
`referred to as the hydraulic conductivity, L; (cm/mm Hg-s).
`The effectiveness of the transluminal
`osmotic pressure differ-
`ence in producing fluid movement across a vessel wall is char-
`acterized by the osmotic reflection coefficient,
`(J is close to 1
`IT;
`for a macromolecule and close to zero for a small molecule
`(16). Thus,
`transport of a molecule across normal or tumor
`vessels is governed by three transport parameters, P, Lp, and (J;
`the surface area for exchange, S; and the transvascular concen-
`tration and pressure gradients.
`Transvascular Transport Parameters. For a macromolecule
`of specified size, charge, configuration,
`and binding constants,
`the transport parameters depend upon the physiological prop-
`erties of the vessel wall (e.g., wall structure, charge). Ultrastruc-
`tural studies of animal and human tumors have shown that
`tumor vessels have wide interendothelial
`junctions,
`a large
`number of fenestrae and transendothelial
`channels formed by
`vesicles, and discontinuous or absent basement membrane (16).
`These characteristics of tumor vessels suggest that
`they should
`have relatively high P and Lp- As a matter of fact, various tissue
`uptake studies have found vascular permeability of tumors to
`be significantly higher than that of skin or muscle (Fig. 2; for a
`review, see Ref. 16). If tumor vessels are indeed "leakier"
`to
`fluid and macromolecules compared to several normal
`tissues,
`what
`leads to their poor extravasation? As discussed below,
`tumors contain regions of high interstitial
`pressure, which
`lowers the fluid extravasation. Since the transvascular
`transport
`of macromolecules under normal conditions occurs primarily
`by convection (16), a decrease in fluid extravasation would lead
`to a decrease in extravasation
`of macromolecules
`(17, 18)?
`Furthermore,
`the average vascular surface area decreases with
`tumor growth; hence one would expect reduced transvascular
`
`Small Tumor
`
`oL.L
`
`---'---
`
`..l----=::::::~
`
`Surrounding
`Middle
`Outer
`Centro I
`Tissue
`Region
`Region
`Region
`(Normal)
`(Tumor)
`(Tumor)
`(Tumor)
`(22)':-'
`Fig. 3. Interstitial pressure gradients in a small versus a large tumor
`Note that the tumor pressure increases with growth while the pressure in the
`surrounding normal
`tissue remains constant. Elevated pressure in the central
`region retards the extravasation of fluid and macromolecules. In addition the
`pressure drop from the center to the periphery leads to an experimentally
`verifiable. radially outward fluid flow.
`
`exchange in large tumors compared to smaller tumors (17).2
`in-
`Transvascular Pressure Gradients. Decreased p- and/or
`creased Pi in tumors has been indirectly demonstrated by several
`investigators working with tumors grown in transparent
`cham-
`bers. By raising venous pressure in the chamber or by loosening
`the chamber, blood flow can be restored in ischemic/necrotic
`in
`tumor areas. Direct measurements
`in sandwich tumors or
`the superficial
`layer of three-dimensional
`tumors have shown
`that on the arterial side vascular pressure does not differ signif-
`icantly between non tumor and tumor vessels, whereas venous
`pressures may be lower in tumor vessels compared to those in
`normal vessels (for a review, see Ref. 7).
`Since the initial work of Young et at. (19), several
`investiga-
`tors have shown that Pi in tumors is significantly higher
`than
`in normal
`tissues (for a review, see Ref. 20). Further,
`as the
`tumor grows, Pi rises up to 30 mm Hg, presumably due to the
`proliferation of tumor cells in a confined space and the absence
`of functioning lymphatic vessels (5, 20). This increase in Pi also
`correlates with a reduction in tumor blood flow and the devel-
`opment of necrosis in a growing tumor (20). Investigations
`of
`intratumor pressure gradients show that the interstitial pressure
`is higher in the center of a tumor and it approaches normal
`physiological pressure towards the periphery [Fig. 3 (20-22)').
`tissues 1rv and 1ri are approximately 20-25 and s-
`In normal
`IS mm Hg, respectively (17,18). Although there are no direct
`measurements of 1ri in tumors, based on high vascular permea-
`bility and high interstitial diffusion coefficient
`in tumors, one
`would expect higher concentration of endogenous plasma pro-
`teins in the tumor
`interstitium than in normal
`interstitium.
`This hypothesis is supported by the dat'in
`the literature (23).
`As a result,
`1ri in tumors may be higher
`than that
`in normal
`tissues, and may lead to reduced osmotic flow.
`in tumors
`is close to zero in the
`As shown in Fig. 3, Pi
`periphery,
`therefore
`the filtration of fluid from vessels, JF,
`would be close to normal. However, as one moves towards the
`center of the tumor,
`the increase in Pi would reduce the extrav-
`asation of fluid, JF. As stated earlier, convective transport of a
`to JF;
`macromolecule
`is proportional
`therefore,
`the rate of
`extravasation of a blood-borne macromolecule would be negli-
`gible in the center of a tumor
`(17, 18). Since transvascular
`
`Z L. T. Baxter and R. K. Jain. Transport of fluid and macromolecules in
`tumors. II. Role of heterogeneous perfusion, manuscript in preparation.
`
`J M. Misiewicz and R. K. Jain. Interstitial pressure gradients in VX2 carci-
`noma, manuscript in preparation.
`816,
`
`4
`
`

`

`TUMOR PHYSIOLOGY AND ANTIBODY DELIVERY
`
`~-D~_O
`
`"
`
`,
`
`~~
`
`°--0_0 -0
`;l
`..............iii
`
`"
`",.
`
`~
`~
`
`-.
`
`'!f. I N
`0 ~ • De xtran
`•
`0 ~ • Ig G
`
`0
`
`0
`
`AI bumin
`
`100
`
`10
`
`0.1
`
`0.01
`
`•
`<,
`NE
`u
`
`~Q
`
`X C
`
`transport by diffusion is negligible for a macromolecule to begin
`with, macromolecular
`extravasation would be very small in the
`high interstitial pressure regions of a tumor. Since high pressure
`regions usually coincide with regions of poor perfusion rate and
`lower vessel surface area,
`leakage of blood borne macromole-
`cules from vessels would be further restricted."
`
`through Interstitial Space
`Transport
`its movement occurs
`Once a macromolecule has extravasated,
`by diffusion and convection through the interstitial
`space. Dif-
`fusion is proportional
`to the concentration
`gradient
`in the
`interstitium,
`and convection is proportional
`to the interstitial
`fluid velocity, u, (cm/s). The latter,
`in turn,
`is proportional
`to
`the pressure gradient
`in the interstitium. The proportionality
`constant which relates diffusive flux to the concentration
`gra-
`dient
`is referred to as the interstitial
`diffusion coefficient, D
`(cm2/s), and the constant which relates u. to the pressure
`gradient
`is referred to as the interstitial hydraulic conductivity,
`K(cm'/mm Hg-s) (for a review, see Ref. 20). Values oftransport
`coefficients D and K are determined by the structure and com-
`position of the interstitial compartment
`as well as the physico-
`chemical properties of the solute molecule. Larger values of
`these parameters
`lead to less hindered movement of fluid and
`macromolecules
`through the interstitium. Similarly,
`large val-
`ues of interstitial pressure and concentration
`gradients
`lead to
`large convective and diffusive fluxes.
`space in
`Interstitial Transport Coefficients. The interstitial
`tumors,
`in general, is very large compared to that in host normal
`tissues (for a review, see Ref. 20). Similar to normal
`tissues, the
`interstitial
`space of tumors
`is composed predominantly
`of a
`collagen and elastic fiber network.
`Interdispcrsed within this
`cross-linked structure are the interstitial
`fluid and macromolec-
`ular constituents
`(polysaccharides) which form a hydrophilic
`gel. While collagen and elastin impart structural
`integrity to a
`tissue, the polysaccharides
`(glycosaminoglycan
`and proteogly-
`cans) are presumably responsible for the resistance to fluid and
`macromolecular motion in the interstitium.
`In several
`tumors
`studied to date, collagen content of tumors is higher than that
`of the host normal
`tissue. On the other hand, hyaluronate
`and
`proteoglycans are, in general, present
`in lower concentrations
`in tumors than in the host normal
`tissue (for a review, see Ref.
`20). The lower concentration of these polysaccharide molecules
`is presumably due to increased activity of lytic enzymes, e.g.,
`hyaluronidase,
`in the tumor
`interstitial
`fluid (for a review, see
`Ref. 5).
`of poly-
`space and low concentrations
`The large interstitial
`saccharides suggest
`that values of K and D should be relatively
`high in tumors. As a matter of fact,
`the data on hydraulic
`conductivity of hepatoma 5123 (24) and the data on effective
`diffusion coefficients of various macromolecules
`in VX2 carci-
`noma (Fig. 4) (25, 26)' support
`this hypothesis. An order of
`magnitude higher values of D and K in tumors
`compared
`compared to several normal
`tissues should favor movement of
`macromolecules
`in the tumor
`interstitium. Then, why do the
`exogenously injected macromolecules not distribute uniformly
`in tumors? As discussed below,
`there arc two reasons for this
`apparent paradox.
`Large Distances in the Interstitium. The time constant,
`r D,
`for a molecule, with diffusion coefficient D, to move by diffusion
`across distance Q is ~pproximately Q'/4D. For diffnsion of IgG
`in tumors (using Dlfrom Fig. 4), Tv is of the order of 1 h for
`
`0.001
`10,000
`
`Molecular Weight
`
`100,000
`
`200,000
`
`Fig. 4. Molecular weight dependence of effective diffusion coefficients, D, of
`dextrans
`(25, 26), albumin (25), and IgG4
`in water
`tumor
`(T), and a
`(W),
`nontumor
`is hindered in both tissues compared to
`tissue (N). Note that transport
`water. Despite higher values of D in tumors compared to in non tumor
`tissues,
`macromolecules
`do not reach uniform concentration in a large tumor for a long
`time due to large diffusion distances.
`
`4 M. A. Clauss andR. K. Jain.
`diffusion coefficients of IgG in
`Interstitial
`normal and neoplastic tissues, submitted for publication.
`
`5 L. T. Baxter and R. K. Jain. Transport
`of fluid and macromolecules
`tumors.
`III. Role of binding and metabolism, manuscript
`in preparation.
`817s
`
`in
`
`5
`
`

`

`Interstitial Fluid Loss from Periphery of the Tumor. It is a
`well known law of physics that fluid flows from a high to a low
`pressure region. As discussed earlier, Pi is high in the center of
`tumors and low in the periphery. Therefore one would expect
`interstitial
`fluid motion from the center of a tumor towards its
`periphery from where it will ooze out
`into the surrounding
`normal
`tissue. Using a tissue isolated tumor preparation, Butler
`et al. (30) measured this fluid loss to be 0.14-0.22 ml/h/g tissue
`in four different
`rat mammary carcinomas. This fluid leakage
`leads to a radially outward interstitial
`fluid velocity of 0.1-0.2
`JLm/s at